summaryrefslogtreecommitdiff
Diffstat
-rwxr-xr-xAUTHORS23
-rwxr-xr-xAndroid.mk89
-rwxr-xr-xAndroid.mk.bak89
-rwxr-xr-xCOPYING340
-rwxr-xr-xCOPYING.LIB482
-rwxr-xr-xCREDITS41
-rwxr-xr-xChangeLog7
-rwxr-xr-xINSTALL237
-rwxr-xr-xMakefile715
-rwxr-xr-xMakefile.am66
-rwxr-xr-xMakefile.in722
-rwxr-xr-xNEWS5
-rwxr-xr-xREADME68
-rwxr-xr-xaclocal.m47723
-rwxr-xr-xautogen.sh22
-rwxr-xr-xcompile142
-rwxr-xr-xconfig.guess1536
-rwxr-xr-xconfig.h361
-rwxr-xr-xconfig.h.in350
-rwxr-xr-xconfig.log3290
-rwxr-xr-xconfig.status1220
-rwxr-xr-xconfig.sub1658
-rwxr-xr-xconfigure25302
-rwxr-xr-xconfigure.ac441
-rwxr-xr-xdepcomp589
-rwxr-xr-xinclude/Makefile.am4
-rwxr-xr-xinclude/Makefile.in503
-rwxr-xr-xinclude/fuse-lite/Makefile.am10
-rwxr-xr-xinclude/fuse-lite/Makefile.in400
-rwxr-xr-xinclude/fuse-lite/fuse.h654
-rwxr-xr-xinclude/fuse-lite/fuse_common.h193
-rwxr-xr-xinclude/fuse-lite/fuse_kernel.h422
-rwxr-xr-xinclude/fuse-lite/fuse_lowlevel.h1375
-rwxr-xr-xinclude/fuse-lite/fuse_lowlevel_compat.h16
-rwxr-xr-xinclude/fuse-lite/fuse_opt.h261
-rwxr-xr-xinclude/ntfs-3g/Makefile.am45
-rwxr-xr-xinclude/ntfs-3g/Makefile.in471
-rwxr-xr-xinclude/ntfs-3g/acls.h199
-rwxr-xr-xinclude/ntfs-3g/attrib.h358
-rwxr-xr-xinclude/ntfs-3g/attrlist.h51
-rwxr-xr-xinclude/ntfs-3g/bitmap.h96
-rwxr-xr-xinclude/ntfs-3g/bootsect.h42
-rwxr-xr-xinclude/ntfs-3g/cache.h115
-rwxr-xr-xinclude/ntfs-3g/collate.h34
-rwxr-xr-xinclude/ntfs-3g/compat.h69
-rwxr-xr-xinclude/ntfs-3g/compress.h39
-rwxr-xr-xinclude/ntfs-3g/debug.h47
-rwxr-xr-xinclude/ntfs-3g/device.h128
-rwxr-xr-xinclude/ntfs-3g/device_io.h77
-rwxr-xr-xinclude/ntfs-3g/dir.h128
-rwxr-xr-xinclude/ntfs-3g/efs.h30
-rwxr-xr-xinclude/ntfs-3g/endians.h203
-rwxr-xr-xinclude/ntfs-3g/index.h167
-rwxr-xr-xinclude/ntfs-3g/inode.h225
-rwxr-xr-xinclude/ntfs-3g/layout.h2661
-rwxr-xr-xinclude/ntfs-3g/lcnalloc.h50
-rwxr-xr-xinclude/ntfs-3g/logfile.h394
-rwxr-xr-xinclude/ntfs-3g/logging.h118
-rwxr-xr-xinclude/ntfs-3g/mft.h132
-rwxr-xr-xinclude/ntfs-3g/misc.h30
-rwxr-xr-xinclude/ntfs-3g/mst.h34
-rwxr-xr-xinclude/ntfs-3g/ntfstime.h131
-rwxr-xr-xinclude/ntfs-3g/object_id.h35
-rwxr-xr-xinclude/ntfs-3g/param.h63
-rwxr-xr-xinclude/ntfs-3g/reparse.h39
-rwxr-xr-xinclude/ntfs-3g/runlist.h89
-rwxr-xr-xinclude/ntfs-3g/security.h353
-rwxr-xr-xinclude/ntfs-3g/support.h85
-rwxr-xr-xinclude/ntfs-3g/types.h124
-rwxr-xr-xinclude/ntfs-3g/unistr.h116
-rwxr-xr-xinclude/ntfs-3g/volume.h277
-rwxr-xr-xinstall-sh519
-rwxr-xr-xlibfuse-lite/Makefile.am28
-rwxr-xr-xlibfuse-lite/Makefile.in572
-rwxr-xr-xlibfuse-lite/fuse.c2789
-rwxr-xr-xlibfuse-lite/fuse_i.h25
-rwxr-xr-xlibfuse-lite/fuse_kern_chan.c96
-rwxr-xr-xlibfuse-lite/fuse_loop.c40
-rwxr-xr-xlibfuse-lite/fuse_lowlevel.c1395
-rwxr-xr-xlibfuse-lite/fuse_misc.h106
-rwxr-xr-xlibfuse-lite/fuse_opt.c368
-rwxr-xr-xlibfuse-lite/fuse_session.c183
-rwxr-xr-xlibfuse-lite/fuse_signals.c73
-rwxr-xr-xlibfuse-lite/fusermount.c772
-rwxr-xr-xlibfuse-lite/helper.c40
-rwxr-xr-xlibfuse-lite/mount.c256
-rwxr-xr-xlibfuse-lite/mount_util.c219
-rwxr-xr-xlibfuse-lite/mount_util.h22
-rwxr-xr-xlibntfs-3g/Makefile.am79
-rwxr-xr-xlibntfs-3g/Makefile.in841
-rwxr-xr-xlibntfs-3g/acls.c4296
-rwxr-xr-xlibntfs-3g/attrib.c5913
-rwxr-xr-xlibntfs-3g/attrlist.c314
-rwxr-xr-xlibntfs-3g/bitmap.c300
-rwxr-xr-xlibntfs-3g/bootsect.c285
-rwxr-xr-xlibntfs-3g/cache.c609
-rwxr-xr-xlibntfs-3g/collate.c271
-rwxr-xr-xlibntfs-3g/compat.c250
-rwxr-xr-xlibntfs-3g/compress.c1431
-rwxr-xr-xlibntfs-3g/debug.c79
-rwxr-xr-xlibntfs-3g/device.c730
-rwxr-xr-xlibntfs-3g/dir.c2582
-rwxr-xr-xlibntfs-3g/efs.c346
-rwxr-xr-xlibntfs-3g/index.c2063
-rwxr-xr-xlibntfs-3g/inode.c1566
-rwxr-xr-xlibntfs-3g/lcnalloc.c735
-rwxr-xr-xlibntfs-3g/libntfs-3g.pc.in10
-rwxr-xr-xlibntfs-3g/libntfs-3g.script.so.in2
-rwxr-xr-xlibntfs-3g/logfile.c737
-rwxr-xr-xlibntfs-3g/logging.c613
-rwxr-xr-xlibntfs-3g/mft.c1909
-rwxr-xr-xlibntfs-3g/misc.c61
-rwxr-xr-xlibntfs-3g/mst.c231
-rwxr-xr-xlibntfs-3g/object_id.c637
-rwxr-xr-xlibntfs-3g/reparse.c1222
-rwxr-xr-xlibntfs-3g/runlist.c2166
-rwxr-xr-xlibntfs-3g/security.c5167
-rwxr-xr-xlibntfs-3g/unistr.c1321
-rwxr-xr-xlibntfs-3g/unix_io.c386
-rwxr-xr-xlibntfs-3g/volume.c1663
-rwxr-xr-xlibntfs-3g/win32_io.c1477
-rwxr-xr-xlibtool7621
-rwxr-xr-xltmain.sh6956
-rwxr-xr-xm4/.keep0
-rwxr-xr-xmissing367
-rwxr-xr-xprog.IAB1071
-rwxr-xr-xprog.IAD5
-rwxr-xr-xprog.IMB466
-rwxr-xr-xprog.IMD2
-rwxr-xr-xprog.PFI2
-rwxr-xr-xprog.PO1
-rwxr-xr-xprog.PR14
-rwxr-xr-xprog.PRI219
-rwxr-xr-xprog.PS979
-rwxr-xr-xprog.SearchResults3
-rwxr-xr-xprog.WK39
-rwxr-xr-xsrc/lowntfs-3g.c4492
-rwxr-xr-xsrc/ntfs-3g.8.in349
-rwxr-xr-xsrc/ntfs-3g.c4461
-rwxr-xr-xsrc/ntfs-3g.probe.8.in81
-rwxr-xr-xsrc/ntfs-3g.probe.c163
-rwxr-xr-xsrc/ntfs-3g.secaudit.8.in171
-rwxr-xr-xsrc/ntfs-3g.usermap.8.in96
-rwxr-xr-xsrc/secaudit.c7176
-rwxr-xr-xsrc/secaudit.h731
-rwxr-xr-xsrc/test.c88
-rwxr-xr-xsrc/usermap.c1356
-rwxr-xr-xstamp-h11
148 files changed, 141286 insertions, 0 deletions
diff --git a/libntfs-3g/acls.c b/libntfs-3g/acls.c
new file mode 100755
index 0000000..11054ba
--- a/dev/null
+++ b/libntfs-3g/acls.c
@@ -0,0 +1,4296 @@
+/**
+ * acls.c - General function to process NTFS ACLs
+ *
+ * This module is part of ntfs-3g library, but may also be
+ * integrated in tools running over Linux or Windows
+ *
+ * Copyright (c) 2007-2009 Jean-Pierre Andre
+ *
+ * This program/include file is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as published
+ * by the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program/include file is distributed in the hope that it will be
+ * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program (in the main directory of the NTFS-3G
+ * distribution in the file COPYING); if not, write to the Free Software
+ * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#ifdef HAVE_CONFIG_H
+ /*
+ * integration into ntfs-3g
+ */
+#include "config.h"
+
+#ifdef HAVE_STDIO_H
+#include <stdio.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifdef HAVE_SYSLOG_H
+#include <syslog.h>
+#endif
+#include <unistd.h>
+#include <pwd.h>
+#include <grp.h>
+
+#include "types.h"
+#include "layout.h"
+#include "security.h"
+#include "acls.h"
+#include "misc.h"
+#else
+
+ /*
+ * integration into secaudit, check whether Win32,
+ * may have to be adapted to compiler or something else
+ */
+
+#ifndef WIN32
+#if defined(__WIN32) | defined(__WIN32__) | defined(WNSC)
+#define WIN32 1
+#endif
+#endif
+
+#include <stdio.h>
+#include <time.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+
+#include <errno.h>
+
+ /*
+ * integration into secaudit/Win32
+ */
+#ifdef WIN32
+#include <fcntl.h>
+#include <windows.h>
+#define __LITTLE_ENDIAN 1234
+#define __BYTE_ORDER __LITTLE_ENDIAN
+#else
+ /*
+ * integration into secaudit/STSC
+ */
+#ifdef STSC
+#include <stat.h>
+#undef __BYTE_ORDER
+#define __BYTE_ORDER __BIG_ENDIAN
+#else
+ /*
+ * integration into secaudit/Linux
+ */
+#include <sys/stat.h>
+#include <endian.h>
+#include <unistd.h>
+#include <dlfcn.h>
+#endif /* STSC */
+#endif /* WIN32 */
+//#include "secaudit.h"
+#endif /* HAVE_CONFIG_H */
+
+/*
+ * A few useful constants
+ */
+
+/*
+ * null SID (S-1-0-0)
+ */
+
+static const char nullsidbytes[] = {
+ 1, /* revision */
+ 1, /* auth count */
+ 0, 0, 0, 0, 0, 0, /* base */
+ 0, 0, 0, 0 /* 1st level */
+ };
+
+static const SID *nullsid = (const SID*)nullsidbytes;
+
+/*
+ * SID for world (S-1-1-0)
+ */
+
+static const char worldsidbytes[] = {
+ 1, /* revision */
+ 1, /* auth count */
+ 0, 0, 0, 0, 0, 1, /* base */
+ 0, 0, 0, 0 /* 1st level */
+} ;
+
+const SID *worldsid = (const SID*)worldsidbytes;
+
+/*
+ * SID for administrator
+ */
+
+static const char adminsidbytes[] = {
+ 1, /* revision */
+ 2, /* auth count */
+ 0, 0, 0, 0, 0, 5, /* base */
+ 32, 0, 0, 0, /* 1st level */
+ 32, 2, 0, 0 /* 2nd level */
+};
+
+const SID *adminsid = (const SID*)adminsidbytes;
+
+/*
+ * SID for system
+ */
+
+static const char systemsidbytes[] = {
+ 1, /* revision */
+ 1, /* auth count */
+ 0, 0, 0, 0, 0, 5, /* base */
+ 18, 0, 0, 0 /* 1st level */
+ };
+
+static const SID *systemsid = (const SID*)systemsidbytes;
+
+/*
+ * SID for generic creator-owner
+ * S-1-3-0
+ */
+
+static const char ownersidbytes[] = {
+ 1, /* revision */
+ 1, /* auth count */
+ 0, 0, 0, 0, 0, 3, /* base */
+ 0, 0, 0, 0 /* 1st level */
+} ;
+
+static const SID *ownersid = (const SID*)ownersidbytes;
+
+/*
+ * SID for generic creator-group
+ * S-1-3-1
+ */
+
+static const char groupsidbytes[] = {
+ 1, /* revision */
+ 1, /* auth count */
+ 0, 0, 0, 0, 0, 3, /* base */
+ 1, 0, 0, 0 /* 1st level */
+} ;
+
+static const SID *groupsid = (const SID*)groupsidbytes;
+
+/*
+ * Determine the size of a SID
+ */
+
+int ntfs_sid_size(const SID * sid)
+{
+ return (sid->sub_authority_count * 4 + 8);
+}
+
+/*
+ * Test whether two SID are equal
+ */
+
+BOOL ntfs_same_sid(const SID *first, const SID *second)
+{
+ int size;
+
+ size = ntfs_sid_size(first);
+ return ((ntfs_sid_size(second) == size)
+ && !memcmp(first, second, size));
+}
+
+/*
+ * Test whether a SID means "world user"
+ * Local users group also recognized as world
+ */
+
+static int is_world_sid(const SID * usid)
+{
+ return (
+ /* check whether S-1-1-0 : world */
+ ((usid->sub_authority_count == 1)
+ && (usid->identifier_authority.high_part == const_cpu_to_be16(0))
+ && (usid->identifier_authority.low_part == const_cpu_to_be32(1))
+ && (usid->sub_authority[0] == const_cpu_to_le32(0)))
+
+ /* check whether S-1-5-32-545 : local user */
+ || ((usid->sub_authority_count == 2)
+ && (usid->identifier_authority.high_part == const_cpu_to_be16(0))
+ && (usid->identifier_authority.low_part == const_cpu_to_be32(5))
+ && (usid->sub_authority[0] == const_cpu_to_le32(32))
+ && (usid->sub_authority[1] == const_cpu_to_le32(545)))
+ );
+}
+
+/*
+ * Test whether a SID means "some user (or group)"
+ * Currently we only check for S-1-5-21... but we should
+ * probably test for other configurations
+ */
+
+BOOL ntfs_is_user_sid(const SID *usid)
+{
+ return ((usid->sub_authority_count == 5)
+ && (usid->identifier_authority.high_part == const_cpu_to_be16(0))
+ && (usid->identifier_authority.low_part == const_cpu_to_be32(5))
+ && (usid->sub_authority[0] == const_cpu_to_le32(21)));
+}
+
+/*
+ * Determine the size of a security attribute
+ * whatever the order of fields
+ */
+
+unsigned int ntfs_attr_size(const char *attr)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pdacl;
+ const ACL *psacl;
+ const SID *psid;
+ unsigned int offdacl;
+ unsigned int offsacl;
+ unsigned int offowner;
+ unsigned int offgroup;
+ unsigned int endsid;
+ unsigned int endacl;
+ unsigned int attrsz;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)attr;
+ /*
+ * First check group, which is the last field in all descriptors
+ * we build, and in most descriptors built by Windows
+ */
+ attrsz = sizeof(SECURITY_DESCRIPTOR_RELATIVE);
+ offgroup = le32_to_cpu(phead->group);
+ if (offgroup >= attrsz) {
+ /* find end of GSID */
+ psid = (const SID*)&attr[offgroup];
+ endsid = offgroup + ntfs_sid_size(psid);
+ if (endsid > attrsz) attrsz = endsid;
+ }
+ offowner = le32_to_cpu(phead->owner);
+ if (offowner >= attrsz) {
+ /* find end of USID */
+ psid = (const SID*)&attr[offowner];
+ endsid = offowner + ntfs_sid_size(psid);
+ attrsz = endsid;
+ }
+ offsacl = le32_to_cpu(phead->sacl);
+ if (offsacl >= attrsz) {
+ /* find end of SACL */
+ psacl = (const ACL*)&attr[offsacl];
+ endacl = offsacl + le16_to_cpu(psacl->size);
+ if (endacl > attrsz)
+ attrsz = endacl;
+ }
+
+
+ /* find end of DACL */
+ offdacl = le32_to_cpu(phead->dacl);
+ if (offdacl >= attrsz) {
+ pdacl = (const ACL*)&attr[offdacl];
+ endacl = offdacl + le16_to_cpu(pdacl->size);
+ if (endacl > attrsz)
+ attrsz = endacl;
+ }
+ return (attrsz);
+}
+
+/*
+ * Do sanity checks on a SID read from storage
+ * (just check revision and number of authorities)
+ */
+
+BOOL ntfs_valid_sid(const SID *sid)
+{
+ return ((sid->revision == SID_REVISION)
+ && (sid->sub_authority_count >= 1)
+ && (sid->sub_authority_count <= 8));
+}
+
+/*
+ * Check whether a SID is acceptable for an implicit
+ * mapping pattern.
+ * It should have been already checked it is a valid user SID.
+ *
+ * The last authority reference has to be >= 1000 (Windows usage)
+ * and <= 0x7fffffff, so that 30 bits from a uid and 30 more bits
+ * from a gid an be inserted with no overflow.
+ */
+
+BOOL ntfs_valid_pattern(const SID *sid)
+{
+ int cnt;
+ u32 auth;
+ le32 leauth;
+
+ cnt = sid->sub_authority_count;
+ leauth = sid->sub_authority[cnt-1];
+ auth = le32_to_cpu(leauth);
+ return ((auth >= 1000) && (auth <= 0x7fffffff));
+}
+
+/*
+ * Compute the uid or gid associated to a SID
+ * through an implicit mapping
+ *
+ * Returns 0 (root) if it does not match pattern
+ */
+
+static u32 findimplicit(const SID *xsid, const SID *pattern, int parity)
+{
+ BIGSID defsid;
+ SID *psid;
+ u32 xid; /* uid or gid */
+ int cnt;
+ u32 carry;
+ le32 leauth;
+ u32 uauth;
+ u32 xlast;
+ u32 rlast;
+
+ memcpy(&defsid,pattern,ntfs_sid_size(pattern));
+ psid = (SID*)&defsid;
+ cnt = psid->sub_authority_count;
+ xid = 0;
+ if (xsid->sub_authority_count == cnt) {
+ psid->sub_authority[cnt-1] = xsid->sub_authority[cnt-1];
+ leauth = xsid->sub_authority[cnt-1];
+ xlast = le32_to_cpu(leauth);
+ leauth = pattern->sub_authority[cnt-1];
+ rlast = le32_to_cpu(leauth);
+
+ if ((xlast > rlast) && !((xlast ^ rlast ^ parity) & 1)) {
+ /* direct check for basic situation */
+ if (ntfs_same_sid(psid,xsid))
+ xid = ((xlast - rlast) >> 1) & 0x3fffffff;
+ else {
+ /*
+ * check whether part of mapping had to be
+ * recorded in a higher level authority
+ */
+ carry = 1;
+ do {
+ leauth = psid->sub_authority[cnt-2];
+ uauth = le32_to_cpu(leauth) + 1;
+ psid->sub_authority[cnt-2]
+ = cpu_to_le32(uauth);
+ } while (!ntfs_same_sid(psid,xsid)
+ && (++carry < 4));
+ if (carry < 4)
+ xid = (((xlast - rlast) >> 1)
+ & 0x3fffffff) | (carry << 30);
+ }
+ }
+ }
+ return (xid);
+}
+
+/*
+ * Find usid mapped to a Linux user
+ * Returns NULL if not found
+ */
+
+const SID *ntfs_find_usid(const struct MAPPING* usermapping,
+ uid_t uid, SID *defusid)
+{
+ const struct MAPPING *p;
+ const SID *sid;
+ le32 leauth;
+ u32 uauth;
+ int cnt;
+
+ if (!uid)
+ sid = adminsid;
+ else {
+ p = usermapping;
+ while (p && p->xid && ((uid_t)p->xid != uid))
+ p = p->next;
+ if (p && !p->xid) {
+ /*
+ * default pattern has been reached :
+ * build an implicit SID according to pattern
+ * (the pattern format was checked while reading
+ * the mapping file)
+ */
+ memcpy(defusid, p->sid, ntfs_sid_size(p->sid));
+ cnt = defusid->sub_authority_count;
+ leauth = defusid->sub_authority[cnt-1];
+ uauth = le32_to_cpu(leauth) + 2*(uid & 0x3fffffff);
+ defusid->sub_authority[cnt-1] = cpu_to_le32(uauth);
+ if (uid & 0xc0000000) {
+ leauth = defusid->sub_authority[cnt-2];
+ uauth = le32_to_cpu(leauth) + ((uid >> 30) & 3);
+ defusid->sub_authority[cnt-2] = cpu_to_le32(uauth);
+ }
+ sid = defusid;
+ } else
+ sid = (p ? p->sid : (const SID*)NULL);
+ }
+ return (sid);
+}
+
+/*
+ * Find Linux group mapped to a gsid
+ * Returns 0 (root) if not found
+ */
+
+const SID *ntfs_find_gsid(const struct MAPPING* groupmapping,
+ gid_t gid, SID *defgsid)
+{
+ const struct MAPPING *p;
+ const SID *sid;
+ le32 leauth;
+ u32 uauth;
+ int cnt;
+
+ if (!gid)
+ sid = adminsid;
+ else {
+ p = groupmapping;
+ while (p && p->xid && ((gid_t)p->xid != gid))
+ p = p->next;
+ if (p && !p->xid) {
+ /*
+ * default pattern has been reached :
+ * build an implicit SID according to pattern
+ * (the pattern format was checked while reading
+ * the mapping file)
+ */
+ memcpy(defgsid, p->sid, ntfs_sid_size(p->sid));
+ cnt = defgsid->sub_authority_count;
+ leauth = defgsid->sub_authority[cnt-1];
+ uauth = le32_to_cpu(leauth) + 2*(gid & 0x3fffffff) + 1;
+ defgsid->sub_authority[cnt-1] = cpu_to_le32(uauth);
+ if (gid & 0xc0000000) {
+ leauth = defgsid->sub_authority[cnt-2];
+ uauth = le32_to_cpu(leauth) + ((gid >> 30) & 3);
+ defgsid->sub_authority[cnt-2] = cpu_to_le32(uauth);
+ }
+ sid = defgsid;
+ } else
+ sid = (p ? p->sid : (const SID*)NULL);
+ }
+ return (sid);
+}
+
+/*
+ * Find Linux owner mapped to a usid
+ * Returns 0 (root) if not found
+ */
+
+uid_t ntfs_find_user(const struct MAPPING* usermapping, const SID *usid)
+{
+ uid_t uid;
+ const struct MAPPING *p;
+
+ p = usermapping;
+ while (p && p->xid && !ntfs_same_sid(usid, p->sid))
+ p = p->next;
+ if (p && !p->xid)
+ /*
+ * No explicit mapping found, try implicit mapping
+ */
+ uid = findimplicit(usid,p->sid,0);
+ else
+ uid = (p ? p->xid : 0);
+ return (uid);
+}
+
+/*
+ * Find Linux group mapped to a gsid
+ * Returns 0 (root) if not found
+ */
+
+gid_t ntfs_find_group(const struct MAPPING* groupmapping, const SID * gsid)
+{
+ gid_t gid;
+ const struct MAPPING *p;
+ int gsidsz;
+
+ gsidsz = ntfs_sid_size(gsid);
+ p = groupmapping;
+ while (p && p->xid && !ntfs_same_sid(gsid, p->sid))
+ p = p->next;
+ if (p && !p->xid)
+ /*
+ * No explicit mapping found, try implicit mapping
+ */
+ gid = findimplicit(gsid,p->sid,1);
+ else
+ gid = (p ? p->xid : 0);
+ return (gid);
+}
+
+/*
+ * Check the validity of the ACEs in a DACL or SACL
+ */
+
+static BOOL valid_acl(const ACL *pacl, unsigned int end)
+{
+ const ACCESS_ALLOWED_ACE *pace;
+ unsigned int offace;
+ unsigned int acecnt;
+ unsigned int acesz;
+ unsigned int nace;
+ BOOL ok;
+
+ ok = TRUE;
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = sizeof(ACL);
+ for (nace = 0; (nace < acecnt) && ok; nace++) {
+ /* be sure the beginning is within range */
+ if ((offace + sizeof(ACCESS_ALLOWED_ACE)) > end)
+ ok = FALSE;
+ else {
+ pace = (const ACCESS_ALLOWED_ACE*)
+ &((const char*)pacl)[offace];
+ acesz = le16_to_cpu(pace->size);
+ if (((offace + acesz) > end)
+ || !ntfs_valid_sid(&pace->sid)
+ || ((ntfs_sid_size(&pace->sid) + 8) != (int)acesz))
+ ok = FALSE;
+ offace += acesz;
+ }
+ }
+ return (ok);
+}
+
+/*
+ * Do sanity checks on security descriptors read from storage
+ * basically, we make sure that every field holds within
+ * allocated storage
+ * Should not be called with a NULL argument
+ * returns TRUE if considered safe
+ * if not, error should be logged by caller
+ */
+
+BOOL ntfs_valid_descr(const char *securattr, unsigned int attrsz)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pdacl;
+ const ACL *psacl;
+ unsigned int offdacl;
+ unsigned int offsacl;
+ unsigned int offowner;
+ unsigned int offgroup;
+ BOOL ok;
+
+ ok = TRUE;
+
+ /*
+ * first check overall size if within allocation range
+ * and a DACL is present
+ * and owner and group SID are valid
+ */
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ offsacl = le32_to_cpu(phead->sacl);
+ offowner = le32_to_cpu(phead->owner);
+ offgroup = le32_to_cpu(phead->group);
+ pdacl = (const ACL*)&securattr[offdacl];
+ psacl = (const ACL*)&securattr[offsacl];
+
+ /*
+ * size check occurs before the above pointers are used
+ *
+ * "DR Watson" standard directory on WinXP has an
+ * old revision and no DACL though SE_DACL_PRESENT is set
+ */
+ if ((attrsz >= sizeof(SECURITY_DESCRIPTOR_RELATIVE))
+ && (ntfs_attr_size(securattr) <= attrsz)
+ && (phead->revision == SECURITY_DESCRIPTOR_REVISION)
+ && (offowner >= sizeof(SECURITY_DESCRIPTOR_RELATIVE))
+ && ((offowner + 2) < attrsz)
+ && (offgroup >= sizeof(SECURITY_DESCRIPTOR_RELATIVE))
+ && ((offgroup + 2) < attrsz)
+ && (!offdacl
+ || ((offdacl >= sizeof(SECURITY_DESCRIPTOR_RELATIVE))
+ && (offdacl < attrsz)))
+ && (!offsacl
+ || ((offsacl >= sizeof(SECURITY_DESCRIPTOR_RELATIVE))
+ && (offsacl < attrsz)))
+ && !(phead->owner & const_cpu_to_le32(3))
+ && !(phead->group & const_cpu_to_le32(3))
+ && !(phead->dacl & const_cpu_to_le32(3))
+ && !(phead->sacl & const_cpu_to_le32(3))
+ && ntfs_valid_sid((const SID*)&securattr[offowner])
+ && ntfs_valid_sid((const SID*)&securattr[offgroup])
+ /*
+ * if there is an ACL, as indicated by offdacl,
+ * require SE_DACL_PRESENT
+ * but "Dr Watson" has SE_DACL_PRESENT though no DACL
+ */
+ && (!offdacl
+ || ((phead->control & SE_DACL_PRESENT)
+ && ((pdacl->revision == ACL_REVISION)
+ || (pdacl->revision == ACL_REVISION_DS))))
+ /* same for SACL */
+ && (!offsacl
+ || ((phead->control & SE_SACL_PRESENT)
+ && ((psacl->revision == ACL_REVISION)
+ || (psacl->revision == ACL_REVISION_DS))))) {
+ /*
+ * Check the DACL and SACL if present
+ */
+ if ((offdacl && !valid_acl(pdacl,attrsz - offdacl))
+ || (offsacl && !valid_acl(psacl,attrsz - offsacl)))
+ ok = FALSE;
+ } else
+ ok = FALSE;
+ return (ok);
+}
+
+/*
+ * Copy the inheritable parts of an ACL
+ *
+ * Returns the size of the new ACL
+ * or zero if nothing is inheritable
+ */
+
+int ntfs_inherit_acl(const ACL *oldacl, ACL *newacl,
+ const SID *usid, const SID *gsid, BOOL fordir)
+{
+ unsigned int src;
+ unsigned int dst;
+ int oldcnt;
+ int newcnt;
+ unsigned int selection;
+ int nace;
+ int acesz;
+ int usidsz;
+ int gsidsz;
+ const ACCESS_ALLOWED_ACE *poldace;
+ ACCESS_ALLOWED_ACE *pnewace;
+
+ usidsz = ntfs_sid_size(usid);
+ gsidsz = ntfs_sid_size(gsid);
+
+ /* ACL header */
+
+ newacl->revision = ACL_REVISION;
+ newacl->alignment1 = 0;
+ newacl->alignment2 = const_cpu_to_le16(0);
+ src = dst = sizeof(ACL);
+
+ selection = (fordir ? CONTAINER_INHERIT_ACE : OBJECT_INHERIT_ACE);
+ newcnt = 0;
+ oldcnt = le16_to_cpu(oldacl->ace_count);
+ for (nace = 0; nace < oldcnt; nace++) {
+ poldace = (const ACCESS_ALLOWED_ACE*)((const char*)oldacl + src);
+ acesz = le16_to_cpu(poldace->size);
+ /* inheritance for access */
+ if (poldace->flags & selection) {
+ pnewace = (ACCESS_ALLOWED_ACE*)
+ ((char*)newacl + dst);
+ memcpy(pnewace,poldace,acesz);
+ /*
+ * Replace generic creator-owner and
+ * creator-group by owner and group
+ */
+ if (ntfs_same_sid(&pnewace->sid, ownersid)) {
+ memcpy(&pnewace->sid, usid, usidsz);
+ acesz = usidsz + 8;
+ pnewace->size = cpu_to_le16(acesz);
+ }
+ if (ntfs_same_sid(&pnewace->sid, groupsid)) {
+ memcpy(&pnewace->sid, gsid, gsidsz);
+ acesz = gsidsz + 8;
+ pnewace->size = cpu_to_le16(acesz);
+ }
+ if (pnewace->mask & GENERIC_ALL) {
+ pnewace->mask &= ~GENERIC_ALL;
+ if (fordir)
+ pnewace->mask |= OWNER_RIGHTS
+ | DIR_READ
+ | DIR_WRITE
+ | DIR_EXEC;
+ else
+ /*
+ * The last flag is not defined for a file,
+ * however Windows sets it, so do the same
+ */
+ pnewace->mask |= OWNER_RIGHTS
+ | FILE_READ
+ | FILE_WRITE
+ | FILE_EXEC
+ | cpu_to_le32(0x40);
+ }
+ /* remove inheritance flags */
+ pnewace->flags &= ~(OBJECT_INHERIT_ACE
+ | CONTAINER_INHERIT_ACE
+ | INHERIT_ONLY_ACE);
+ dst += acesz;
+ newcnt++;
+ }
+ /* inheritance for further inheritance */
+ if (fordir
+ && (poldace->flags
+ & (CONTAINER_INHERIT_ACE | OBJECT_INHERIT_ACE))) {
+ pnewace = (ACCESS_ALLOWED_ACE*)
+ ((char*)newacl + dst);
+ memcpy(pnewace,poldace,acesz);
+ /*
+ * Replace generic creator-owner and
+ * creator-group by owner and group
+ */
+ if (ntfs_same_sid(&pnewace->sid, ownersid)) {
+ memcpy(&pnewace->sid, usid, usidsz);
+ acesz = usidsz + 8;
+ }
+ if (ntfs_same_sid(&pnewace->sid, groupsid)) {
+ memcpy(&pnewace->sid, gsid, gsidsz);
+ acesz = gsidsz + 8;
+ }
+ dst += acesz;
+ newcnt++;
+ }
+ src += acesz;
+ }
+ /*
+ * Adjust header if something was inherited
+ */
+ if (dst > sizeof(ACL)) {
+ newacl->ace_count = cpu_to_le16(newcnt);
+ newacl->size = cpu_to_le16(dst);
+ } else
+ dst = 0;
+ return (dst);
+}
+
+#if POSIXACLS
+
+/*
+ * Do sanity checks on a Posix descriptor
+ * Should not be called with a NULL argument
+ * returns TRUE if considered safe
+ * if not, error should be logged by caller
+ */
+
+BOOL ntfs_valid_posix(const struct POSIX_SECURITY *pxdesc)
+{
+ const struct POSIX_ACL *pacl;
+ int i;
+ BOOL ok;
+ u16 tag;
+ u32 id;
+ int perms;
+ struct {
+ u16 previous;
+ u32 previousid;
+ u16 tagsset;
+ mode_t mode;
+ int owners;
+ int groups;
+ int others;
+ } checks[2], *pchk;
+
+ for (i=0; i<2; i++) {
+ checks[i].mode = 0;
+ checks[i].tagsset = 0;
+ checks[i].owners = 0;
+ checks[i].groups = 0;
+ checks[i].others = 0;
+ checks[i].previous = 0;
+ checks[i].previousid = 0;
+ }
+ ok = TRUE;
+ pacl = &pxdesc->acl;
+ /*
+ * header (strict for now)
+ */
+ if ((pacl->version != POSIX_VERSION)
+ || (pacl->flags != 0)
+ || (pacl->filler != 0))
+ ok = FALSE;
+ /*
+ * Reject multiple owner, group or other
+ * but do not require them to be present
+ * Also check the ACEs are in correct order
+ * which implies there is no duplicates
+ */
+ for (i=0; i<pxdesc->acccnt + pxdesc->defcnt; i++) {
+ if (i >= pxdesc->firstdef)
+ pchk = &checks[1];
+ else
+ pchk = &checks[0];
+ perms = pacl->ace[i].perms;
+ tag = pacl->ace[i].tag;
+ pchk->tagsset |= tag;
+ id = pacl->ace[i].id;
+ if (perms & ~7) ok = FALSE;
+ if ((tag < pchk->previous)
+ || ((tag == pchk->previous)
+ && (id <= pchk->previousid)))
+ ok = FALSE;
+ pchk->previous = tag;
+ pchk->previousid = id;
+ switch (tag) {
+ case POSIX_ACL_USER_OBJ :
+ if (pchk->owners++)
+ ok = FALSE;
+ if (id != (u32)-1)
+ ok = FALSE;
+ pchk->mode |= perms << 6;
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ if (pchk->groups++)
+ ok = FALSE;
+ if (id != (u32)-1)
+ ok = FALSE;
+ pchk->mode = (pchk->mode & 07707) | (perms << 3);
+ break;
+ case POSIX_ACL_OTHER :
+ if (pchk->others++)
+ ok = FALSE;
+ if (id != (u32)-1)
+ ok = FALSE;
+ pchk->mode |= perms;
+ break;
+ case POSIX_ACL_USER :
+ case POSIX_ACL_GROUP :
+ if (id == (u32)-1)
+ ok = FALSE;
+ break;
+ case POSIX_ACL_MASK :
+ if (id != (u32)-1)
+ ok = FALSE;
+ pchk->mode = (pchk->mode & 07707) | (perms << 3);
+ break;
+ default :
+ ok = FALSE;
+ break;
+ }
+ }
+ if ((pxdesc->acccnt > 0)
+ && ((checks[0].owners != 1) || (checks[0].groups != 1)
+ || (checks[0].others != 1)))
+ ok = FALSE;
+ /* do not check owner, group or other are present in */
+ /* the default ACL, Windows does not necessarily set them */
+ /* descriptor */
+ if (pxdesc->defcnt && (pxdesc->acccnt > pxdesc->firstdef))
+ ok = FALSE;
+ if ((pxdesc->acccnt < 0) || (pxdesc->defcnt < 0))
+ ok = FALSE;
+ /* check mode, unless null or no tag set */
+ if (pxdesc->mode
+ && checks[0].tagsset
+ && (checks[0].mode != (pxdesc->mode & 0777)))
+ ok = FALSE;
+ /* check tagsset */
+ if (pxdesc->tagsset != checks[0].tagsset)
+ ok = FALSE;
+ return (ok);
+}
+
+/*
+ * Set standard header data into a Posix ACL
+ * The mode argument should provide the 3 upper bits of target mode
+ */
+
+static mode_t posix_header(struct POSIX_SECURITY *pxdesc, mode_t basemode)
+{
+ mode_t mode;
+ u16 tagsset;
+ struct POSIX_ACE *pace;
+ int i;
+
+ mode = basemode & 07000;
+ tagsset = 0;
+ for (i=0; i<pxdesc->acccnt; i++) {
+ pace = &pxdesc->acl.ace[i];
+ tagsset |= pace->tag;
+ switch(pace->tag) {
+ case POSIX_ACL_USER_OBJ :
+ mode |= (pace->perms & 7) << 6;
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ case POSIX_ACL_MASK :
+ mode = (mode & 07707) | ((pace->perms & 7) << 3);
+ break;
+ case POSIX_ACL_OTHER :
+ mode |= pace->perms & 7;
+ break;
+ default :
+ break;
+ }
+ }
+ pxdesc->tagsset = tagsset;
+ pxdesc->mode = mode;
+ pxdesc->acl.version = POSIX_VERSION;
+ pxdesc->acl.flags = 0;
+ pxdesc->acl.filler = 0;
+ return (mode);
+}
+
+/*
+ * Sort ACEs in a Posix ACL
+ * This is useful for always getting reusable converted ACLs,
+ * it also helps in merging ACEs.
+ * Repeated tag+id are allowed and not merged here.
+ *
+ * Tags should be in ascending sequence and for a repeatable tag
+ * ids should be in ascending sequence.
+ */
+
+void ntfs_sort_posix(struct POSIX_SECURITY *pxdesc)
+{
+ struct POSIX_ACL *pacl;
+ struct POSIX_ACE ace;
+ int i;
+ int offs;
+ BOOL done;
+ u16 tag;
+ u16 previous;
+ u32 id;
+ u32 previousid;
+
+
+ /*
+ * Check sequencing of tag+id in access ACE's
+ */
+ pacl = &pxdesc->acl;
+ do {
+ done = TRUE;
+ previous = pacl->ace[0].tag;
+ previousid = pacl->ace[0].id;
+ for (i=1; i<pxdesc->acccnt; i++) {
+ tag = pacl->ace[i].tag;
+ id = pacl->ace[i].id;
+
+ if ((tag < previous)
+ || ((tag == previous) && (id < previousid))) {
+ done = FALSE;
+ memcpy(&ace,&pacl->ace[i-1],sizeof(struct POSIX_ACE));
+ memcpy(&pacl->ace[i-1],&pacl->ace[i],sizeof(struct POSIX_ACE));
+ memcpy(&pacl->ace[i],&ace,sizeof(struct POSIX_ACE));
+ } else {
+ previous = tag;
+ previousid = id;
+ }
+ }
+ } while (!done);
+ /*
+ * Same for default ACEs
+ */
+ do {
+ done = TRUE;
+ if ((pxdesc->defcnt) > 1) {
+ offs = pxdesc->firstdef;
+ previous = pacl->ace[offs].tag;
+ previousid = pacl->ace[offs].id;
+ for (i=offs+1; i<offs+pxdesc->defcnt; i++) {
+ tag = pacl->ace[i].tag;
+ id = pacl->ace[i].id;
+
+ if ((tag < previous)
+ || ((tag == previous) && (id < previousid))) {
+ done = FALSE;
+ memcpy(&ace,&pacl->ace[i-1],sizeof(struct POSIX_ACE));
+ memcpy(&pacl->ace[i-1],&pacl->ace[i],sizeof(struct POSIX_ACE));
+ memcpy(&pacl->ace[i],&ace,sizeof(struct POSIX_ACE));
+ } else {
+ previous = tag;
+ previousid = id;
+ }
+ }
+ }
+ } while (!done);
+}
+
+/*
+ * Merge a new mode into a Posix descriptor
+ * The Posix descriptor is not reallocated, its size is unchanged
+ *
+ * returns 0 if ok
+ */
+
+int ntfs_merge_mode_posix(struct POSIX_SECURITY *pxdesc, mode_t mode)
+{
+ int i;
+ BOOL maskfound;
+ struct POSIX_ACE *pace;
+ int todo;
+
+ maskfound = FALSE;
+ todo = POSIX_ACL_USER_OBJ | POSIX_ACL_GROUP_OBJ | POSIX_ACL_OTHER;
+ for (i=pxdesc->acccnt-1; i>=0; i--) {
+ pace = &pxdesc->acl.ace[i];
+ switch(pace->tag) {
+ case POSIX_ACL_USER_OBJ :
+ pace->perms = (mode >> 6) & 7;
+ todo &= ~POSIX_ACL_USER_OBJ;
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ if (!maskfound)
+ pace->perms = (mode >> 3) & 7;
+ todo &= ~POSIX_ACL_GROUP_OBJ;
+ break;
+ case POSIX_ACL_MASK :
+ pace->perms = (mode >> 3) & 7;
+ maskfound = TRUE;
+ break;
+ case POSIX_ACL_OTHER :
+ pace->perms = mode & 7;
+ todo &= ~POSIX_ACL_OTHER;
+ break;
+ default :
+ break;
+ }
+ }
+ pxdesc->mode = mode;
+ return (todo ? -1 : 0);
+}
+
+/*
+ * Replace an access or default Posix ACL
+ * The resulting ACL is checked for validity
+ *
+ * Returns a new ACL or NULL if there is a problem
+ */
+
+struct POSIX_SECURITY *ntfs_replace_acl(const struct POSIX_SECURITY *oldpxdesc,
+ const struct POSIX_ACL *newacl, int count, BOOL deflt)
+{
+ struct POSIX_SECURITY *newpxdesc;
+ size_t newsize;
+ int offset;
+ int oldoffset;
+ int i;
+
+ if (deflt)
+ newsize = sizeof(struct POSIX_SECURITY)
+ + (oldpxdesc->acccnt + count)*sizeof(struct POSIX_ACE);
+ else
+ newsize = sizeof(struct POSIX_SECURITY)
+ + (oldpxdesc->defcnt + count)*sizeof(struct POSIX_ACE);
+ newpxdesc = (struct POSIX_SECURITY*)malloc(newsize);
+ if (newpxdesc) {
+ if (deflt) {
+ offset = oldpxdesc->acccnt;
+ newpxdesc->acccnt = oldpxdesc->acccnt;
+ newpxdesc->defcnt = count;
+ newpxdesc->firstdef = offset;
+ /* copy access ACEs */
+ for (i=0; i<newpxdesc->acccnt; i++)
+ newpxdesc->acl.ace[i] = oldpxdesc->acl.ace[i];
+ /* copy default ACEs */
+ for (i=0; i<count; i++)
+ newpxdesc->acl.ace[i + offset] = newacl->ace[i];
+ } else {
+ offset = count;
+ newpxdesc->acccnt = count;
+ newpxdesc->defcnt = oldpxdesc->defcnt;
+ newpxdesc->firstdef = count;
+ /* copy access ACEs */
+ for (i=0; i<count; i++)
+ newpxdesc->acl.ace[i] = newacl->ace[i];
+ /* copy default ACEs */
+ oldoffset = oldpxdesc->firstdef;
+ for (i=0; i<newpxdesc->defcnt; i++)
+ newpxdesc->acl.ace[i + offset] = oldpxdesc->acl.ace[i + oldoffset];
+ }
+ /* assume special flags unchanged */
+ posix_header(newpxdesc, oldpxdesc->mode);
+ if (!ntfs_valid_posix(newpxdesc)) {
+ /* do not log, this is an application error */
+ free(newpxdesc);
+ newpxdesc = (struct POSIX_SECURITY*)NULL;
+ errno = EINVAL;
+ }
+ } else
+ errno = ENOMEM;
+ return (newpxdesc);
+}
+
+/*
+ * Build an inherited Posix descriptor from parent
+ * descriptor (if any) restricted to creation mode
+ *
+ * Returns the inherited descriptor or NULL if there is a problem
+ */
+
+struct POSIX_SECURITY *ntfs_build_inherited_posix(
+ const struct POSIX_SECURITY *pxdesc, mode_t mode,
+ mode_t mask, BOOL isdir)
+{
+ struct POSIX_SECURITY *pydesc;
+ struct POSIX_ACE *pyace;
+ int count;
+ int defcnt;
+ int size;
+ int i;
+ s16 tagsset;
+
+ if (pxdesc && pxdesc->defcnt) {
+ if (isdir)
+ count = 2*pxdesc->defcnt + 3;
+ else
+ count = pxdesc->defcnt + 3;
+ } else
+ count = 3;
+ pydesc = (struct POSIX_SECURITY*)malloc(
+ sizeof(struct POSIX_SECURITY) + count*sizeof(struct POSIX_ACE));
+ if (pydesc) {
+ /*
+ * Copy inherited tags and adapt perms
+ * Use requested mode, ignoring umask
+ * (not possible with older versions of fuse)
+ */
+ tagsset = 0;
+ defcnt = (pxdesc ? pxdesc->defcnt : 0);
+ for (i=defcnt-1; i>=0; i--) {
+ pyace = &pydesc->acl.ace[i];
+ *pyace = pxdesc->acl.ace[pxdesc->firstdef + i];
+ switch (pyace->tag) {
+ case POSIX_ACL_USER_OBJ :
+ pyace->perms &= (mode >> 6) & 7;
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ if (!(tagsset & POSIX_ACL_MASK))
+ pyace->perms &= (mode >> 3) & 7;
+ break;
+ case POSIX_ACL_OTHER :
+ pyace->perms &= mode & 7;
+ break;
+ case POSIX_ACL_MASK :
+ pyace->perms &= (mode >> 3) & 7;
+ break;
+ default :
+ break;
+ }
+ tagsset |= pyace->tag;
+ }
+ pydesc->acccnt = defcnt;
+ /*
+ * If some standard tags were missing, append them from mode
+ * and sort the list
+ * Here we have to use the umask'ed mode
+ */
+ if (~tagsset & (POSIX_ACL_USER_OBJ
+ | POSIX_ACL_GROUP_OBJ | POSIX_ACL_OTHER)) {
+ i = defcnt;
+ /* owner was missing */
+ if (!(tagsset & POSIX_ACL_USER_OBJ)) {
+ pyace = &pydesc->acl.ace[i];
+ pyace->tag = POSIX_ACL_USER_OBJ;
+ pyace->id = -1;
+ pyace->perms = ((mode & ~mask) >> 6) & 7;
+ tagsset |= POSIX_ACL_USER_OBJ;
+ i++;
+ }
+ /* owning group was missing */
+ if (!(tagsset & POSIX_ACL_GROUP_OBJ)) {
+ pyace = &pydesc->acl.ace[i];
+ pyace->tag = POSIX_ACL_GROUP_OBJ;
+ pyace->id = -1;
+ pyace->perms = ((mode & ~mask) >> 3) & 7;
+ tagsset |= POSIX_ACL_GROUP_OBJ;
+ i++;
+ }
+ /* other was missing */
+ if (!(tagsset & POSIX_ACL_OTHER)) {
+ pyace = &pydesc->acl.ace[i];
+ pyace->tag = POSIX_ACL_OTHER;
+ pyace->id = -1;
+ pyace->perms = mode & ~mask & 7;
+ tagsset |= POSIX_ACL_OTHER;
+ i++;
+ }
+ pydesc->acccnt = i;
+ pydesc->firstdef = i;
+ pydesc->defcnt = 0;
+ ntfs_sort_posix(pydesc);
+ }
+
+ /*
+ * append as a default ACL if a directory
+ */
+ pydesc->firstdef = pydesc->acccnt;
+ if (defcnt && isdir) {
+ size = sizeof(struct POSIX_ACE)*defcnt;
+ memcpy(&pydesc->acl.ace[pydesc->firstdef],
+ &pxdesc->acl.ace[pxdesc->firstdef],size);
+ pydesc->defcnt = defcnt;
+ } else {
+ pydesc->defcnt = 0;
+ }
+ /* assume special bits are not inherited */
+ posix_header(pydesc, mode & 07000);
+ if (!ntfs_valid_posix(pydesc)) {
+ ntfs_log_error("Error building an inherited Posix desc\n");
+ errno = EIO;
+ free(pydesc);
+ pydesc = (struct POSIX_SECURITY*)NULL;
+ }
+ } else
+ errno = ENOMEM;
+ return (pydesc);
+}
+
+static int merge_lists_posix(struct POSIX_ACE *targetace,
+ const struct POSIX_ACE *firstace,
+ const struct POSIX_ACE *secondace,
+ int firstcnt, int secondcnt)
+{
+ int k;
+
+ k = 0;
+ /*
+ * No list is exhausted :
+ * if same tag+id in both list :
+ * ignore ACE from second list
+ * else take the one with smaller tag+id
+ */
+ while ((firstcnt > 0) && (secondcnt > 0))
+ if ((firstace->tag == secondace->tag)
+ && (firstace->id == secondace->id)) {
+ secondace++;
+ secondcnt--;
+ } else
+ if ((firstace->tag < secondace->tag)
+ || ((firstace->tag == secondace->tag)
+ && (firstace->id < secondace->id))) {
+ targetace->tag = firstace->tag;
+ targetace->id = firstace->id;
+ targetace->perms = firstace->perms;
+ firstace++;
+ targetace++;
+ firstcnt--;
+ k++;
+ } else {
+ targetace->tag = secondace->tag;
+ targetace->id = secondace->id;
+ targetace->perms = secondace->perms;
+ secondace++;
+ targetace++;
+ secondcnt--;
+ k++;
+ }
+ /*
+ * One list is exhausted, copy the other one
+ */
+ while (firstcnt > 0) {
+ targetace->tag = firstace->tag;
+ targetace->id = firstace->id;
+ targetace->perms = firstace->perms;
+ firstace++;
+ targetace++;
+ firstcnt--;
+ k++;
+ }
+ while (secondcnt > 0) {
+ targetace->tag = secondace->tag;
+ targetace->id = secondace->id;
+ targetace->perms = secondace->perms;
+ secondace++;
+ targetace++;
+ secondcnt--;
+ k++;
+ }
+ return (k);
+}
+
+/*
+ * Merge two Posix ACLs
+ * The input ACLs have to be adequately sorted
+ *
+ * Returns the merged ACL, which is allocated and has to be freed by caller,
+ * or NULL if failed
+ */
+
+struct POSIX_SECURITY *ntfs_merge_descr_posix(const struct POSIX_SECURITY *first,
+ const struct POSIX_SECURITY *second)
+{
+ struct POSIX_SECURITY *pxdesc;
+ struct POSIX_ACE *targetace;
+ const struct POSIX_ACE *firstace;
+ const struct POSIX_ACE *secondace;
+ size_t size;
+ int k;
+
+ size = sizeof(struct POSIX_SECURITY)
+ + (first->acccnt + first->defcnt
+ + second->acccnt + second->defcnt)*sizeof(struct POSIX_ACE);
+ pxdesc = (struct POSIX_SECURITY*)malloc(size);
+ if (pxdesc) {
+ /*
+ * merge access ACEs
+ */
+ firstace = first->acl.ace;
+ secondace = second->acl.ace;
+ targetace = pxdesc->acl.ace;
+ k = merge_lists_posix(targetace,firstace,secondace,
+ first->acccnt,second->acccnt);
+ pxdesc->acccnt = k;
+ /*
+ * merge default ACEs
+ */
+ pxdesc->firstdef = k;
+ firstace = &first->acl.ace[first->firstdef];
+ secondace = &second->acl.ace[second->firstdef];
+ targetace = &pxdesc->acl.ace[k];
+ k = merge_lists_posix(targetace,firstace,secondace,
+ first->defcnt,second->defcnt);
+ pxdesc->defcnt = k;
+ /*
+ * build header
+ */
+ pxdesc->acl.version = POSIX_VERSION;
+ pxdesc->acl.flags = 0;
+ pxdesc->acl.filler = 0;
+ pxdesc->mode = 0;
+ pxdesc->tagsset = 0;
+ } else
+ errno = ENOMEM;
+ return (pxdesc);
+}
+
+struct BUILD_CONTEXT {
+ BOOL isdir;
+ BOOL adminowns;
+ BOOL groupowns;
+ u16 selfuserperms;
+ u16 selfgrpperms;
+ u16 grpperms;
+ u16 othperms;
+ u16 mask;
+ u16 designates;
+ u16 withmask;
+ u16 rootspecial;
+} ;
+
+
+
+static BOOL build_user_denials(ACL *pacl,
+ const SID *usid, struct MAPPING* const mapping[],
+ ACE_FLAGS flags, const struct POSIX_ACE *pxace,
+ struct BUILD_CONTEXT *pset)
+{
+ BIGSID defsid;
+ ACCESS_ALLOWED_ACE *pdace;
+ const SID *sid;
+ int sidsz;
+ int pos;
+ int acecnt;
+ le32 grants;
+ le32 denials;
+ u16 perms;
+ u16 mixperms;
+ u16 tag;
+ BOOL rejected;
+ BOOL rootuser;
+ BOOL avoidmask;
+
+ rejected = FALSE;
+ tag = pxace->tag;
+ perms = pxace->perms;
+ rootuser = FALSE;
+ pos = le16_to_cpu(pacl->size);
+ acecnt = le16_to_cpu(pacl->ace_count);
+ avoidmask = (pset->mask == (POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X))
+ && ((pset->designates && pset->withmask)
+ || (!pset->designates && !pset->withmask));
+ if (tag == POSIX_ACL_USER_OBJ) {
+ sid = usid;
+ sidsz = ntfs_sid_size(sid);
+ grants = OWNER_RIGHTS;
+ } else {
+ if (pxace->id) {
+ sid = NTFS_FIND_USID(mapping[MAPUSERS],
+ pxace->id, (SID*)&defsid);
+ grants = WORLD_RIGHTS;
+ } else {
+ sid = adminsid;
+ rootuser = TRUE;
+ grants = WORLD_RIGHTS & ~ROOT_OWNER_UNMARK;
+ }
+ if (sid) {
+ sidsz = ntfs_sid_size(sid);
+ /*
+ * Insert denial of complement of mask for
+ * each designated user (except root)
+ * WRITE_OWNER is inserted so that
+ * the mask can be identified
+ */
+ if (!avoidmask && !rootuser) {
+ denials = WRITE_OWNER;
+ pdace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ if (pset->isdir) {
+ if (!(pset->mask & POSIX_PERM_X))
+ denials |= DIR_EXEC;
+ if (!(pset->mask & POSIX_PERM_W))
+ denials |= DIR_WRITE;
+ if (!(pset->mask & POSIX_PERM_R))
+ denials |= DIR_READ;
+ } else {
+ if (!(pset->mask & POSIX_PERM_X))
+ denials |= FILE_EXEC;
+ if (!(pset->mask & POSIX_PERM_W))
+ denials |= FILE_WRITE;
+ if (!(pset->mask & POSIX_PERM_R))
+ denials |= FILE_READ;
+ }
+ if (rootuser)
+ grants &= ~ROOT_OWNER_UNMARK;
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = flags;
+ pdace->size = cpu_to_le16(sidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt++;
+ }
+ } else
+ rejected = TRUE;
+ }
+ if (!rejected) {
+ if (pset->isdir) {
+ if (perms & POSIX_PERM_X)
+ grants |= DIR_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= DIR_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= DIR_READ;
+ } else {
+ if (perms & POSIX_PERM_X)
+ grants |= FILE_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= FILE_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= FILE_READ;
+ }
+
+ /* a possible ACE to deny owner what he/she would */
+ /* induely get from administrator, group or world */
+ /* unless owner is administrator or group */
+
+ denials = const_cpu_to_le32(0);
+ pdace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ if (!pset->adminowns && !rootuser) {
+ if (!pset->groupowns) {
+ mixperms = pset->grpperms | pset->othperms;
+ if (tag == POSIX_ACL_USER_OBJ)
+ mixperms |= pset->selfuserperms;
+ if (pset->isdir) {
+ if (mixperms & POSIX_PERM_X)
+ denials |= DIR_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= DIR_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= DIR_READ;
+ } else {
+ if (mixperms & POSIX_PERM_X)
+ denials |= FILE_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= FILE_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= FILE_READ;
+ }
+ } else {
+ mixperms = ~pset->grpperms & pset->othperms;
+ if (tag == POSIX_ACL_USER_OBJ)
+ mixperms |= pset->selfuserperms;
+ if (pset->isdir) {
+ if (mixperms & POSIX_PERM_X)
+ denials |= DIR_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= DIR_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= DIR_READ;
+ } else {
+ if (mixperms & POSIX_PERM_X)
+ denials |= FILE_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= FILE_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= FILE_READ;
+ }
+ }
+ denials &= ~grants;
+ if (denials) {
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = flags;
+ pdace->size = cpu_to_le16(sidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt++;
+ }
+ }
+ }
+ pacl->size = cpu_to_le16(pos);
+ pacl->ace_count = cpu_to_le16(acecnt);
+ return (!rejected);
+}
+
+static BOOL build_user_grants(ACL *pacl,
+ const SID *usid, struct MAPPING* const mapping[],
+ ACE_FLAGS flags, const struct POSIX_ACE *pxace,
+ struct BUILD_CONTEXT *pset)
+{
+ BIGSID defsid;
+ ACCESS_ALLOWED_ACE *pgace;
+ const SID *sid;
+ int sidsz;
+ int pos;
+ int acecnt;
+ le32 grants;
+ u16 perms;
+ u16 tag;
+ BOOL rejected;
+ BOOL rootuser;
+
+ rejected = FALSE;
+ tag = pxace->tag;
+ perms = pxace->perms;
+ rootuser = FALSE;
+ pos = le16_to_cpu(pacl->size);
+ acecnt = le16_to_cpu(pacl->ace_count);
+ if (tag == POSIX_ACL_USER_OBJ) {
+ sid = usid;
+ sidsz = ntfs_sid_size(sid);
+ grants = OWNER_RIGHTS;
+ } else {
+ if (pxace->id) {
+ sid = NTFS_FIND_USID(mapping[MAPUSERS],
+ pxace->id, (SID*)&defsid);
+ if (sid)
+ sidsz = ntfs_sid_size(sid);
+ else
+ rejected = TRUE;
+ grants = WORLD_RIGHTS;
+ } else {
+ sid = adminsid;
+ sidsz = ntfs_sid_size(sid);
+ rootuser = TRUE;
+ grants = WORLD_RIGHTS & ~ROOT_OWNER_UNMARK;
+ }
+ }
+ if (!rejected) {
+ if (pset->isdir) {
+ if (perms & POSIX_PERM_X)
+ grants |= DIR_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= DIR_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= DIR_READ;
+ } else {
+ if (perms & POSIX_PERM_X)
+ grants |= FILE_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= FILE_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= FILE_READ;
+ }
+ if (rootuser)
+ grants &= ~ROOT_OWNER_UNMARK;
+ pgace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->size = cpu_to_le16(sidsz + 8);
+ pgace->flags = flags;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt = le16_to_cpu(pacl->ace_count) + 1;
+ pacl->ace_count = cpu_to_le16(acecnt);
+ pacl->size = cpu_to_le16(pos);
+ }
+ return (!rejected);
+}
+
+
+ /* a grant ACE for group */
+ /* unless group-obj has the same rights as world */
+ /* but present if group is owner or owner is administrator */
+ /* this ACE will be inserted after denials for group */
+
+static BOOL build_group_denials_grant(ACL *pacl,
+ const SID *gsid, struct MAPPING* const mapping[],
+ ACE_FLAGS flags, const struct POSIX_ACE *pxace,
+ struct BUILD_CONTEXT *pset)
+{
+ BIGSID defsid;
+ ACCESS_ALLOWED_ACE *pdace;
+ ACCESS_ALLOWED_ACE *pgace;
+ const SID *sid;
+ int sidsz;
+ int pos;
+ int acecnt;
+ le32 grants;
+ le32 denials;
+ u16 perms;
+ u16 mixperms;
+ u16 tag;
+ BOOL avoidmask;
+ BOOL rootgroup;
+ BOOL rejected;
+
+ rejected = FALSE;
+ tag = pxace->tag;
+ perms = pxace->perms;
+ pos = le16_to_cpu(pacl->size);
+ acecnt = le16_to_cpu(pacl->ace_count);
+ rootgroup = FALSE;
+ avoidmask = (pset->mask == (POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X))
+ && ((pset->designates && pset->withmask)
+ || (!pset->designates && !pset->withmask));
+ if (tag == POSIX_ACL_GROUP_OBJ)
+ sid = gsid;
+ else
+ if (pxace->id)
+ sid = NTFS_FIND_GSID(mapping[MAPGROUPS],
+ pxace->id, (SID*)&defsid);
+ else {
+ sid = adminsid;
+ rootgroup = TRUE;
+ }
+ if (sid) {
+ sidsz = ntfs_sid_size(sid);
+ /*
+ * Insert denial of complement of mask for
+ * each group
+ * WRITE_OWNER is inserted so that
+ * the mask can be identified
+ * Note : this mask may lead on Windows to
+ * deny rights to administrators belonging
+ * to some user group
+ */
+ if ((!avoidmask && !rootgroup)
+ || (pset->rootspecial
+ && (tag == POSIX_ACL_GROUP_OBJ))) {
+ denials = WRITE_OWNER;
+ pdace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ if (pset->isdir) {
+ if (!(pset->mask & POSIX_PERM_X))
+ denials |= DIR_EXEC;
+ if (!(pset->mask & POSIX_PERM_W))
+ denials |= DIR_WRITE;
+ if (!(pset->mask & POSIX_PERM_R))
+ denials |= DIR_READ;
+ } else {
+ if (!(pset->mask & POSIX_PERM_X))
+ denials |= FILE_EXEC;
+ if (!(pset->mask & POSIX_PERM_W))
+ denials |= FILE_WRITE;
+ if (!(pset->mask & POSIX_PERM_R))
+ denials |= FILE_READ;
+ }
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = flags;
+ pdace->size = cpu_to_le16(sidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt++;
+ }
+ } else
+ rejected = TRUE;
+ if (!rejected
+ && (pset->adminowns
+ || pset->groupowns
+ || avoidmask
+ || rootgroup
+ || (perms != pset->othperms))) {
+ grants = WORLD_RIGHTS;
+ if (rootgroup)
+ grants &= ~ROOT_GROUP_UNMARK;
+ if (pset->isdir) {
+ if (perms & POSIX_PERM_X)
+ grants |= DIR_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= DIR_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= DIR_READ;
+ } else {
+ if (perms & POSIX_PERM_X)
+ grants |= FILE_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= FILE_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= FILE_READ;
+ }
+
+ /* a possible ACE to deny group what it would get from world */
+ /* or administrator, unless owner is administrator or group */
+
+ denials = const_cpu_to_le32(0);
+ pdace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ if (!pset->adminowns
+ && !pset->groupowns
+ && !rootgroup) {
+ mixperms = pset->othperms;
+ if (tag == POSIX_ACL_GROUP_OBJ)
+ mixperms |= pset->selfgrpperms;
+ if (pset->isdir) {
+ if (mixperms & POSIX_PERM_X)
+ denials |= DIR_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= DIR_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= DIR_READ;
+ } else {
+ if (mixperms & POSIX_PERM_X)
+ denials |= FILE_EXEC;
+ if (mixperms & POSIX_PERM_W)
+ denials |= FILE_WRITE;
+ if (mixperms & POSIX_PERM_R)
+ denials |= FILE_READ;
+ }
+ denials &= ~(grants | OWNER_RIGHTS);
+ if (denials) {
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = flags;
+ pdace->size = cpu_to_le16(sidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt++;
+ }
+ }
+
+ /* now insert grants to group if more than world */
+ if (pset->adminowns
+ || pset->groupowns
+ || (avoidmask && (pset->designates || pset->withmask))
+ || (perms & ~pset->othperms)
+ || (pset->rootspecial
+ && (tag == POSIX_ACL_GROUP_OBJ))
+ || (tag == POSIX_ACL_GROUP)) {
+ if (rootgroup)
+ grants &= ~ROOT_GROUP_UNMARK;
+ pgace = (ACCESS_DENIED_ACE*)&((char*)pacl)[pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = flags;
+ pgace->size = cpu_to_le16(sidsz + 8);
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, sid, sidsz);
+ pos += sidsz + 8;
+ acecnt++;
+ }
+ }
+ pacl->size = cpu_to_le16(pos);
+ pacl->ace_count = cpu_to_le16(acecnt);
+ return (!rejected);
+}
+
+
+/*
+ * Build an ACL composed of several ACE's
+ * returns size of ACL or zero if failed
+ *
+ * Three schemes are defined :
+ *
+ * 1) if root is neither owner nor group up to 7 ACE's are set up :
+ * - denials to owner (preventing grants to world or group to apply)
+ * + mask denials to designated user (unless mask allows all)
+ * + denials to designated user
+ * - grants to owner (always present - first grant)
+ * + grants to designated user
+ * + mask denial to group (unless mask allows all)
+ * - denials to group (preventing grants to world to apply)
+ * - grants to group (unless group has no more than world rights)
+ * + mask denials to designated group (unless mask allows all)
+ * + grants to designated group
+ * + denials to designated group
+ * - grants to world (unless none)
+ * - full privileges to administrator, always present
+ * - full privileges to system, always present
+ *
+ * The same scheme is applied for Posix ACLs, with the mask represented
+ * as denials prepended to grants for designated users and groups
+ *
+ * This is inspired by an Internet Draft from Marius Aamodt Eriksen
+ * for mapping NFSv4 ACLs to Posix ACLs (draft-ietf-nfsv4-acl-mapping-00.txt)
+ * More recent versions of the draft (draft-ietf-nfsv4-acl-mapping-05.txt)
+ * are not followed, as they ignore the Posix mask and lead to
+ * loss of compatibility with Linux implementations on other fs.
+ *
+ * Note that denials to group are located after grants to owner.
+ * This only occurs in the unfrequent situation where world
+ * has more rights than group and cannot be avoided if owner and other
+ * have some common right which is denied to group (eg for mode 745
+ * executing has to be denied to group, but not to owner or world).
+ * This rare situation is processed by Windows correctly, but
+ * Windows utilities may want to change the order, with a
+ * consequence of applying the group denials to the Windows owner.
+ * The interpretation on Linux is not affected by the order change.
+ *
+ * 2) if root is either owner or group, two problems arise :
+ * - granting full rights to administrator (as needed to transpose
+ * to Windows rights bypassing granting to root) would imply
+ * Linux permissions to always be seen as rwx, no matter the chmod
+ * - there is no different SID to separate an administrator owner
+ * from an administrator group. Hence Linux permissions for owner
+ * would always be similar to permissions to group.
+ *
+ * as a work-around, up to 5 ACE's are set up if owner or group :
+ * - grants to owner, always present at first position
+ * - grants to group, always present
+ * - grants to world, unless none
+ * - full privileges to administrator, always present
+ * - full privileges to system, always present
+ *
+ * On Windows, these ACE's are processed normally, though they
+ * are redundant (owner, group and administrator are the same,
+ * as a consequence any denials would damage administrator rights)
+ * but on Linux, privileges to administrator are ignored (they
+ * are not needed as root has always full privileges), and
+ * neither grants to group are applied to owner, nor grants to
+ * world are applied to owner or group.
+ *
+ * 3) finally a similar situation arises when group is owner (they
+ * have the same SID), but is not root.
+ * In this situation up to 6 ACE's are set up :
+ *
+ * - denials to owner (preventing grants to world to apply)
+ * - grants to owner (always present)
+ * - grants to group (unless groups has same rights as world)
+ * - grants to world (unless none)
+ * - full privileges to administrator, always present
+ * - full privileges to system, always present
+ *
+ * On Windows, these ACE's are processed normally, though they
+ * are redundant (as owner and group are the same), but this has
+ * no impact on administrator rights
+ *
+ * Special flags (S_ISVTX, S_ISGID, S_ISUID) :
+ * an extra null ACE is inserted to hold these flags, using
+ * the same conventions as cygwin.
+ *
+ */
+
+static int buildacls_posix(struct MAPPING* const mapping[],
+ char *secattr, int offs, const struct POSIX_SECURITY *pxdesc,
+ int isdir, const SID *usid, const SID *gsid)
+{
+ struct BUILD_CONTEXT aceset[2], *pset;
+ BOOL adminowns;
+ BOOL groupowns;
+ ACL *pacl;
+ ACCESS_ALLOWED_ACE *pgace;
+ ACCESS_ALLOWED_ACE *pdace;
+ const struct POSIX_ACE *pxace;
+ BOOL ok;
+ mode_t mode;
+ u16 tag;
+ u16 perms;
+ ACE_FLAGS flags;
+ int pos;
+ int i;
+ int k;
+ BIGSID defsid;
+ const SID *sid;
+ int acecnt;
+ int usidsz;
+ int gsidsz;
+ int wsidsz;
+ int asidsz;
+ int ssidsz;
+ int nsidsz;
+ le32 grants;
+
+ usidsz = ntfs_sid_size(usid);
+ gsidsz = ntfs_sid_size(gsid);
+ wsidsz = ntfs_sid_size(worldsid);
+ asidsz = ntfs_sid_size(adminsid);
+ ssidsz = ntfs_sid_size(systemsid);
+ mode = pxdesc->mode;
+ /* adminowns and groupowns are used for both lists */
+ adminowns = ntfs_same_sid(usid, adminsid)
+ || ntfs_same_sid(gsid, adminsid);
+ groupowns = !adminowns && ntfs_same_sid(usid, gsid);
+
+ ok = TRUE;
+
+ /* ACL header */
+ pacl = (ACL*)&secattr[offs];
+ pacl->revision = ACL_REVISION;
+ pacl->alignment1 = 0;
+ pacl->size = cpu_to_le16(sizeof(ACL) + usidsz + 8);
+ pacl->ace_count = const_cpu_to_le16(0);
+ pacl->alignment2 = const_cpu_to_le16(0);
+
+ /*
+ * Determine what is allowed to some group or world
+ * to prevent designated users or other groups to get
+ * rights from groups or world
+ * Do the same if owner and group appear as designated
+ * user or group
+ * Also get global mask
+ */
+ for (k=0; k<2; k++) {
+ pset = &aceset[k];
+ pset->selfuserperms = 0;
+ pset->selfgrpperms = 0;
+ pset->grpperms = 0;
+ pset->othperms = 0;
+ pset->mask = (POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X);
+ pset->designates = 0;
+ pset->withmask = 0;
+ pset->rootspecial = 0;
+ pset->adminowns = adminowns;
+ pset->groupowns = groupowns;
+ pset->isdir = isdir;
+ }
+
+ for (i=pxdesc->acccnt+pxdesc->defcnt-1; i>=0; i--) {
+ if (i >= pxdesc->acccnt) {
+ pset = &aceset[1];
+ pxace = &pxdesc->acl.ace[i + pxdesc->firstdef - pxdesc->acccnt];
+ } else {
+ pset = &aceset[0];
+ pxace = &pxdesc->acl.ace[i];
+ }
+ switch (pxace->tag) {
+ case POSIX_ACL_USER :
+ pset->designates++;
+ if (pxace->id) {
+ sid = NTFS_FIND_USID(mapping[MAPUSERS],
+ pxace->id, (SID*)&defsid);
+ if (sid && ntfs_same_sid(sid,usid))
+ pset->selfuserperms |= pxace->perms;
+ } else
+ /* root as designated user is processed apart */
+ pset->rootspecial = TRUE;
+ break;
+ case POSIX_ACL_GROUP :
+ pset->designates++;
+ if (pxace->id) {
+ sid = NTFS_FIND_GSID(mapping[MAPUSERS],
+ pxace->id, (SID*)&defsid);
+ if (sid && ntfs_same_sid(sid,gsid))
+ pset->selfgrpperms |= pxace->perms;
+ } else
+ /* root as designated group is processed apart */
+ pset->rootspecial = TRUE;
+ /* fall through */
+ case POSIX_ACL_GROUP_OBJ :
+ pset->grpperms |= pxace->perms;
+ break;
+ case POSIX_ACL_OTHER :
+ pset->othperms = pxace->perms;
+ break;
+ case POSIX_ACL_MASK :
+ pset->withmask++;
+ pset->mask = pxace->perms;
+ default :
+ break;
+ }
+ }
+
+if (pxdesc->defcnt && (pxdesc->firstdef != pxdesc->acccnt)) {
+ntfs_log_error("** error : access and default not consecutive\n");
+return (0);
+}
+ /*
+ * First insert all denials for owner and each
+ * designated user (with mask if needed)
+ */
+
+ pacl->ace_count = const_cpu_to_le16(0);
+ pacl->size = const_cpu_to_le16(sizeof(ACL));
+ for (i=0; (i<(pxdesc->acccnt + pxdesc->defcnt)) && ok; i++) {
+ if (i >= pxdesc->acccnt) {
+ flags = INHERIT_ONLY_ACE
+ | OBJECT_INHERIT_ACE | CONTAINER_INHERIT_ACE;
+ pset = &aceset[1];
+ pxace = &pxdesc->acl.ace[i + pxdesc->firstdef - pxdesc->acccnt];
+ } else {
+ if (pxdesc->defcnt)
+ flags = NO_PROPAGATE_INHERIT_ACE;
+ else
+ flags = (isdir ? DIR_INHERITANCE
+ : FILE_INHERITANCE);
+ pset = &aceset[0];
+ pxace = &pxdesc->acl.ace[i];
+ }
+ tag = pxace->tag;
+ perms = pxace->perms;
+ switch (tag) {
+
+ /* insert denial ACEs for each owner or allowed user */
+
+ case POSIX_ACL_USER :
+ case POSIX_ACL_USER_OBJ :
+
+ ok = build_user_denials(pacl,
+ usid, mapping, flags, pxace, pset);
+ break;
+ default :
+ break;
+ }
+ }
+
+ /*
+ * for directories, insert a world execution denial
+ * inherited to plain files.
+ * This is to prevent Windows from granting execution
+ * of files through inheritance from parent directory
+ */
+
+ if (isdir && ok) {
+ pos = le16_to_cpu(pacl->size);
+ pdace = (ACCESS_DENIED_ACE*) &secattr[offs + pos];
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = INHERIT_ONLY_ACE | OBJECT_INHERIT_ACE;
+ pdace->size = cpu_to_le16(wsidsz + 8);
+ pdace->mask = FILE_EXEC;
+ memcpy((char*)&pdace->sid, worldsid, wsidsz);
+ pos += wsidsz + 8;
+ acecnt = le16_to_cpu(pacl->ace_count) + 1;
+ pacl->ace_count = cpu_to_le16(acecnt);
+ pacl->size = cpu_to_le16(pos);
+ }
+
+ /*
+ * now insert (if needed)
+ * - grants to owner and designated users
+ * - mask and denials for all groups
+ * - grants to other
+ */
+
+ for (i=0; (i<(pxdesc->acccnt + pxdesc->defcnt)) && ok; i++) {
+ if (i >= pxdesc->acccnt) {
+ flags = INHERIT_ONLY_ACE
+ | OBJECT_INHERIT_ACE | CONTAINER_INHERIT_ACE;
+ pset = &aceset[1];
+ pxace = &pxdesc->acl.ace[i + pxdesc->firstdef - pxdesc->acccnt];
+ } else {
+ if (pxdesc->defcnt)
+ flags = NO_PROPAGATE_INHERIT_ACE;
+ else
+ flags = (isdir ? DIR_INHERITANCE
+ : FILE_INHERITANCE);
+ pset = &aceset[0];
+ pxace = &pxdesc->acl.ace[i];
+ }
+ tag = pxace->tag;
+ perms = pxace->perms;
+ switch (tag) {
+
+ /* ACE for each owner or allowed user */
+
+ case POSIX_ACL_USER :
+ case POSIX_ACL_USER_OBJ :
+ ok = build_user_grants(pacl,usid,
+ mapping,flags,pxace,pset);
+ break;
+
+ case POSIX_ACL_GROUP :
+ case POSIX_ACL_GROUP_OBJ :
+
+ /* denials and grants for groups */
+
+ ok = build_group_denials_grant(pacl,gsid,
+ mapping,flags,pxace,pset);
+ break;
+
+ case POSIX_ACL_OTHER :
+
+ /* grants for other users */
+
+ pos = le16_to_cpu(pacl->size);
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ grants = WORLD_RIGHTS;
+ if (isdir) {
+ if (perms & POSIX_PERM_X)
+ grants |= DIR_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= DIR_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= DIR_READ;
+ } else {
+ if (perms & POSIX_PERM_X)
+ grants |= FILE_EXEC;
+ if (perms & POSIX_PERM_W)
+ grants |= FILE_WRITE;
+ if (perms & POSIX_PERM_R)
+ grants |= FILE_READ;
+ }
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = flags;
+ pgace->size = cpu_to_le16(wsidsz + 8);
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, worldsid, wsidsz);
+ pos += wsidsz + 8;
+ acecnt = le16_to_cpu(pacl->ace_count) + 1;
+ pacl->ace_count = cpu_to_le16(acecnt);
+ pacl->size = cpu_to_le16(pos);
+ break;
+ }
+ }
+
+ if (!ok) {
+ errno = EINVAL;
+ pos = 0;
+ } else {
+ /* an ACE for administrators */
+ /* always full access */
+
+ pos = le16_to_cpu(pacl->size);
+ acecnt = le16_to_cpu(pacl->ace_count);
+ if (isdir)
+ flags = OBJECT_INHERIT_ACE
+ | CONTAINER_INHERIT_ACE;
+ else
+ flags = NO_PROPAGATE_INHERIT_ACE;
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = flags;
+ pgace->size = cpu_to_le16(asidsz + 8);
+ grants = OWNER_RIGHTS | FILE_READ | FILE_WRITE | FILE_EXEC;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, adminsid, asidsz);
+ pos += asidsz + 8;
+ acecnt++;
+
+ /* an ACE for system (needed ?) */
+ /* always full access */
+
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = flags;
+ pgace->size = cpu_to_le16(ssidsz + 8);
+ grants = OWNER_RIGHTS | FILE_READ | FILE_WRITE | FILE_EXEC;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, systemsid, ssidsz);
+ pos += ssidsz + 8;
+ acecnt++;
+
+ /* a null ACE to hold special flags */
+ /* using the same representation as cygwin */
+
+ if (mode & (S_ISVTX | S_ISGID | S_ISUID)) {
+ nsidsz = ntfs_sid_size(nullsid);
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = NO_PROPAGATE_INHERIT_ACE;
+ pgace->size = cpu_to_le16(nsidsz + 8);
+ grants = const_cpu_to_le32(0);
+ if (mode & S_ISUID)
+ grants |= FILE_APPEND_DATA;
+ if (mode & S_ISGID)
+ grants |= FILE_WRITE_DATA;
+ if (mode & S_ISVTX)
+ grants |= FILE_READ_DATA;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, nullsid, nsidsz);
+ pos += nsidsz + 8;
+ acecnt++;
+ }
+
+ /* fix ACL header */
+ pacl->size = cpu_to_le16(pos);
+ pacl->ace_count = cpu_to_le16(acecnt);
+ }
+ return (ok ? pos : 0);
+}
+
+#endif /* POSIXACLS */
+
+static int buildacls(char *secattr, int offs, mode_t mode, int isdir,
+ const SID * usid, const SID * gsid)
+{
+ ACL *pacl;
+ ACCESS_ALLOWED_ACE *pgace;
+ ACCESS_ALLOWED_ACE *pdace;
+ BOOL adminowns;
+ BOOL groupowns;
+ ACE_FLAGS gflags;
+ int pos;
+ int acecnt;
+ int usidsz;
+ int gsidsz;
+ int wsidsz;
+ int asidsz;
+ int ssidsz;
+ int nsidsz;
+ le32 grants;
+ le32 denials;
+
+ usidsz = ntfs_sid_size(usid);
+ gsidsz = ntfs_sid_size(gsid);
+ wsidsz = ntfs_sid_size(worldsid);
+ asidsz = ntfs_sid_size(adminsid);
+ ssidsz = ntfs_sid_size(systemsid);
+ adminowns = ntfs_same_sid(usid, adminsid)
+ || ntfs_same_sid(gsid, adminsid);
+ groupowns = !adminowns && ntfs_same_sid(usid, gsid);
+
+ /* ACL header */
+ pacl = (ACL*)&secattr[offs];
+ pacl->revision = ACL_REVISION;
+ pacl->alignment1 = 0;
+ pacl->size = cpu_to_le16(sizeof(ACL) + usidsz + 8);
+ pacl->ace_count = const_cpu_to_le16(1);
+ pacl->alignment2 = const_cpu_to_le16(0);
+ pos = sizeof(ACL);
+ acecnt = 0;
+
+ /* compute a grant ACE for owner */
+ /* this ACE will be inserted after denial for owner */
+
+ grants = OWNER_RIGHTS;
+ if (isdir) {
+ gflags = DIR_INHERITANCE;
+ if (mode & S_IXUSR)
+ grants |= DIR_EXEC;
+ if (mode & S_IWUSR)
+ grants |= DIR_WRITE;
+ if (mode & S_IRUSR)
+ grants |= DIR_READ;
+ } else {
+ gflags = FILE_INHERITANCE;
+ if (mode & S_IXUSR)
+ grants |= FILE_EXEC;
+ if (mode & S_IWUSR)
+ grants |= FILE_WRITE;
+ if (mode & S_IRUSR)
+ grants |= FILE_READ;
+ }
+
+ /* a possible ACE to deny owner what he/she would */
+ /* induely get from administrator, group or world */
+ /* unless owner is administrator or group */
+
+ denials = const_cpu_to_le32(0);
+ pdace = (ACCESS_DENIED_ACE*) &secattr[offs + pos];
+ if (!adminowns) {
+ if (!groupowns) {
+ if (isdir) {
+ pdace->flags = DIR_INHERITANCE;
+ if (mode & (S_IXGRP | S_IXOTH))
+ denials |= DIR_EXEC;
+ if (mode & (S_IWGRP | S_IWOTH))
+ denials |= DIR_WRITE;
+ if (mode & (S_IRGRP | S_IROTH))
+ denials |= DIR_READ;
+ } else {
+ pdace->flags = FILE_INHERITANCE;
+ if (mode & (S_IXGRP | S_IXOTH))
+ denials |= FILE_EXEC;
+ if (mode & (S_IWGRP | S_IWOTH))
+ denials |= FILE_WRITE;
+ if (mode & (S_IRGRP | S_IROTH))
+ denials |= FILE_READ;
+ }
+ } else {
+ if (isdir) {
+ pdace->flags = DIR_INHERITANCE;
+ if ((mode & S_IXOTH) && !(mode & S_IXGRP))
+ denials |= DIR_EXEC;
+ if ((mode & S_IWOTH) && !(mode & S_IWGRP))
+ denials |= DIR_WRITE;
+ if ((mode & S_IROTH) && !(mode & S_IRGRP))
+ denials |= DIR_READ;
+ } else {
+ pdace->flags = FILE_INHERITANCE;
+ if ((mode & S_IXOTH) && !(mode & S_IXGRP))
+ denials |= FILE_EXEC;
+ if ((mode & S_IWOTH) && !(mode & S_IWGRP))
+ denials |= FILE_WRITE;
+ if ((mode & S_IROTH) && !(mode & S_IRGRP))
+ denials |= FILE_READ;
+ }
+ }
+ denials &= ~grants;
+ if (denials) {
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->size = cpu_to_le16(usidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, usid, usidsz);
+ pos += usidsz + 8;
+ acecnt++;
+ }
+ }
+ /*
+ * for directories, a world execution denial
+ * inherited to plain files
+ */
+
+ if (isdir) {
+ pdace = (ACCESS_DENIED_ACE*) &secattr[offs + pos];
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->flags = INHERIT_ONLY_ACE | OBJECT_INHERIT_ACE;
+ pdace->size = cpu_to_le16(wsidsz + 8);
+ pdace->mask = FILE_EXEC;
+ memcpy((char*)&pdace->sid, worldsid, wsidsz);
+ pos += wsidsz + 8;
+ acecnt++;
+ }
+
+
+ /* now insert grants to owner */
+ pgace = (ACCESS_ALLOWED_ACE*) &secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->size = cpu_to_le16(usidsz + 8);
+ pgace->flags = gflags;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, usid, usidsz);
+ pos += usidsz + 8;
+ acecnt++;
+
+ /* a grant ACE for group */
+ /* unless group has the same rights as world */
+ /* but present if group is owner or owner is administrator */
+ /* this ACE will be inserted after denials for group */
+
+ if (adminowns
+ || groupowns
+ || (((mode >> 3) ^ mode) & 7)) {
+ grants = WORLD_RIGHTS;
+ if (isdir) {
+ gflags = DIR_INHERITANCE;
+ if (mode & S_IXGRP)
+ grants |= DIR_EXEC;
+ if (mode & S_IWGRP)
+ grants |= DIR_WRITE;
+ if (mode & S_IRGRP)
+ grants |= DIR_READ;
+ } else {
+ gflags = FILE_INHERITANCE;
+ if (mode & S_IXGRP)
+ grants |= FILE_EXEC;
+ if (mode & S_IWGRP)
+ grants |= FILE_WRITE;
+ if (mode & S_IRGRP)
+ grants |= FILE_READ;
+ }
+
+ /* a possible ACE to deny group what it would get from world */
+ /* or administrator, unless owner is administrator or group */
+
+ denials = const_cpu_to_le32(0);
+ pdace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ if (!adminowns && !groupowns) {
+ if (isdir) {
+ pdace->flags = DIR_INHERITANCE;
+ if (mode & S_IXOTH)
+ denials |= DIR_EXEC;
+ if (mode & S_IWOTH)
+ denials |= DIR_WRITE;
+ if (mode & S_IROTH)
+ denials |= DIR_READ;
+ } else {
+ pdace->flags = FILE_INHERITANCE;
+ if (mode & S_IXOTH)
+ denials |= FILE_EXEC;
+ if (mode & S_IWOTH)
+ denials |= FILE_WRITE;
+ if (mode & S_IROTH)
+ denials |= FILE_READ;
+ }
+ denials &= ~(grants | OWNER_RIGHTS);
+ if (denials) {
+ pdace->type = ACCESS_DENIED_ACE_TYPE;
+ pdace->size = cpu_to_le16(gsidsz + 8);
+ pdace->mask = denials;
+ memcpy((char*)&pdace->sid, gsid, gsidsz);
+ pos += gsidsz + 8;
+ acecnt++;
+ }
+ }
+
+ if (adminowns
+ || groupowns
+ || ((mode >> 3) & ~mode & 7)) {
+ /* now insert grants to group */
+ /* if more rights than other */
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = gflags;
+ pgace->size = cpu_to_le16(gsidsz + 8);
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, gsid, gsidsz);
+ pos += gsidsz + 8;
+ acecnt++;
+ }
+ }
+
+ /* an ACE for world users */
+
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ grants = WORLD_RIGHTS;
+ if (isdir) {
+ pgace->flags = DIR_INHERITANCE;
+ if (mode & S_IXOTH)
+ grants |= DIR_EXEC;
+ if (mode & S_IWOTH)
+ grants |= DIR_WRITE;
+ if (mode & S_IROTH)
+ grants |= DIR_READ;
+ } else {
+ pgace->flags = FILE_INHERITANCE;
+ if (mode & S_IXOTH)
+ grants |= FILE_EXEC;
+ if (mode & S_IWOTH)
+ grants |= FILE_WRITE;
+ if (mode & S_IROTH)
+ grants |= FILE_READ;
+ }
+ pgace->size = cpu_to_le16(wsidsz + 8);
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, worldsid, wsidsz);
+ pos += wsidsz + 8;
+ acecnt++;
+
+ /* an ACE for administrators */
+ /* always full access */
+
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ if (isdir)
+ pgace->flags = DIR_INHERITANCE;
+ else
+ pgace->flags = FILE_INHERITANCE;
+ pgace->size = cpu_to_le16(asidsz + 8);
+ grants = OWNER_RIGHTS | FILE_READ | FILE_WRITE | FILE_EXEC;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, adminsid, asidsz);
+ pos += asidsz + 8;
+ acecnt++;
+
+ /* an ACE for system (needed ?) */
+ /* always full access */
+
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ if (isdir)
+ pgace->flags = DIR_INHERITANCE;
+ else
+ pgace->flags = FILE_INHERITANCE;
+ pgace->size = cpu_to_le16(ssidsz + 8);
+ grants = OWNER_RIGHTS | FILE_READ | FILE_WRITE | FILE_EXEC;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, systemsid, ssidsz);
+ pos += ssidsz + 8;
+ acecnt++;
+
+ /* a null ACE to hold special flags */
+ /* using the same representation as cygwin */
+
+ if (mode & (S_ISVTX | S_ISGID | S_ISUID)) {
+ nsidsz = ntfs_sid_size(nullsid);
+ pgace = (ACCESS_ALLOWED_ACE*)&secattr[offs + pos];
+ pgace->type = ACCESS_ALLOWED_ACE_TYPE;
+ pgace->flags = NO_PROPAGATE_INHERIT_ACE;
+ pgace->size = cpu_to_le16(nsidsz + 8);
+ grants = const_cpu_to_le32(0);
+ if (mode & S_ISUID)
+ grants |= FILE_APPEND_DATA;
+ if (mode & S_ISGID)
+ grants |= FILE_WRITE_DATA;
+ if (mode & S_ISVTX)
+ grants |= FILE_READ_DATA;
+ pgace->mask = grants;
+ memcpy((char*)&pgace->sid, nullsid, nsidsz);
+ pos += nsidsz + 8;
+ acecnt++;
+ }
+
+ /* fix ACL header */
+ pacl->size = cpu_to_le16(pos);
+ pacl->ace_count = cpu_to_le16(acecnt);
+ return (pos);
+}
+
+#if POSIXACLS
+
+/*
+ * Build a full security descriptor from a Posix ACL
+ * returns descriptor in allocated memory, must free() after use
+ */
+
+char *ntfs_build_descr_posix(struct MAPPING* const mapping[],
+ struct POSIX_SECURITY *pxdesc,
+ int isdir, const SID *usid, const SID *gsid)
+{
+ int newattrsz;
+ SECURITY_DESCRIPTOR_RELATIVE *pnhead;
+ char *newattr;
+ int aclsz;
+ int usidsz;
+ int gsidsz;
+ int wsidsz;
+ int asidsz;
+ int ssidsz;
+ int k;
+
+ usidsz = ntfs_sid_size(usid);
+ gsidsz = ntfs_sid_size(gsid);
+ wsidsz = ntfs_sid_size(worldsid);
+ asidsz = ntfs_sid_size(adminsid);
+ ssidsz = ntfs_sid_size(systemsid);
+
+ /* allocate enough space for the new security attribute */
+ newattrsz = sizeof(SECURITY_DESCRIPTOR_RELATIVE) /* header */
+ + usidsz + gsidsz /* usid and gsid */
+ + sizeof(ACL) /* acl header */
+ + 2*(8 + usidsz) /* two possible ACE for user */
+ + 3*(8 + gsidsz) /* three possible ACE for group and mask */
+ + 8 + wsidsz /* one ACE for world */
+ + 8 + asidsz /* one ACE for admin */
+ + 8 + ssidsz; /* one ACE for system */
+ if (isdir) /* a world denial for directories */
+ newattrsz += 8 + wsidsz;
+ if (pxdesc->mode & 07000) /* a NULL ACE for special modes */
+ newattrsz += 8 + ntfs_sid_size(nullsid);
+ /* account for non-owning users and groups */
+ for (k=0; k<pxdesc->acccnt; k++) {
+ if ((pxdesc->acl.ace[k].tag == POSIX_ACL_USER)
+ || (pxdesc->acl.ace[k].tag == POSIX_ACL_GROUP))
+ newattrsz += 3*40; /* fixme : maximum size */
+ }
+ /* account for default ACE's */
+ newattrsz += 2*40*pxdesc->defcnt; /* fixme : maximum size */
+ newattr = (char*)ntfs_malloc(newattrsz);
+ if (newattr) {
+ /* build the main header part */
+ pnhead = (SECURITY_DESCRIPTOR_RELATIVE*)newattr;
+ pnhead->revision = SECURITY_DESCRIPTOR_REVISION;
+ pnhead->alignment = 0;
+ /*
+ * The flag SE_DACL_PROTECTED prevents the ACL
+ * to be changed in an inheritance after creation
+ */
+ pnhead->control = SE_DACL_PRESENT | SE_DACL_PROTECTED
+ | SE_SELF_RELATIVE;
+ /*
+ * Windows prefers ACL first, do the same to
+ * get the same hash value and avoid duplication
+ */
+ /* build permissions */
+ aclsz = buildacls_posix(mapping,newattr,
+ sizeof(SECURITY_DESCRIPTOR_RELATIVE),
+ pxdesc, isdir, usid, gsid);
+ if (aclsz && ((int)(sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz + gsidsz) <= newattrsz)) {
+ /* append usid and gsid */
+ memcpy(&newattr[sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz], usid, usidsz);
+ memcpy(&newattr[sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz], gsid, gsidsz);
+ /* positions of ACL, USID and GSID into header */
+ pnhead->owner =
+ cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz);
+ pnhead->group =
+ cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz);
+ pnhead->sacl = const_cpu_to_le32(0);
+ pnhead->dacl =
+ const_cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE));
+ } else {
+ /* ACL failure (errno set) or overflow */
+ free(newattr);
+ newattr = (char*)NULL;
+ if (aclsz) {
+ /* hope error was detected before overflowing */
+ ntfs_log_error("Security descriptor is longer than expected\n");
+ errno = EIO;
+ }
+ }
+ } else
+ errno = ENOMEM;
+ return (newattr);
+}
+
+#endif /* POSIXACLS */
+
+/*
+ * Build a full security descriptor
+ * returns descriptor in allocated memory, must free() after use
+ */
+
+char *ntfs_build_descr(mode_t mode,
+ int isdir, const SID * usid, const SID * gsid)
+{
+ int newattrsz;
+ SECURITY_DESCRIPTOR_RELATIVE *pnhead;
+ char *newattr;
+ int aclsz;
+ int usidsz;
+ int gsidsz;
+ int wsidsz;
+ int asidsz;
+ int ssidsz;
+
+ usidsz = ntfs_sid_size(usid);
+ gsidsz = ntfs_sid_size(gsid);
+ wsidsz = ntfs_sid_size(worldsid);
+ asidsz = ntfs_sid_size(adminsid);
+ ssidsz = ntfs_sid_size(systemsid);
+
+ /* allocate enough space for the new security attribute */
+ newattrsz = sizeof(SECURITY_DESCRIPTOR_RELATIVE) /* header */
+ + usidsz + gsidsz /* usid and gsid */
+ + sizeof(ACL) /* acl header */
+ + 2*(8 + usidsz) /* two possible ACE for user */
+ + 2*(8 + gsidsz) /* two possible ACE for group */
+ + 8 + wsidsz /* one ACE for world */
+ + 8 + asidsz /* one ACE for admin */
+ + 8 + ssidsz; /* one ACE for system */
+ if (isdir) /* a world denial for directories */
+ newattrsz += 8 + wsidsz;
+ if (mode & 07000) /* a NULL ACE for special modes */
+ newattrsz += 8 + ntfs_sid_size(nullsid);
+ newattr = (char*)ntfs_malloc(newattrsz);
+ if (newattr) {
+ /* build the main header part */
+ pnhead = (SECURITY_DESCRIPTOR_RELATIVE*) newattr;
+ pnhead->revision = SECURITY_DESCRIPTOR_REVISION;
+ pnhead->alignment = 0;
+ /*
+ * The flag SE_DACL_PROTECTED prevents the ACL
+ * to be changed in an inheritance after creation
+ */
+ pnhead->control = SE_DACL_PRESENT | SE_DACL_PROTECTED
+ | SE_SELF_RELATIVE;
+ /*
+ * Windows prefers ACL first, do the same to
+ * get the same hash value and avoid duplication
+ */
+ /* build permissions */
+ aclsz = buildacls(newattr,
+ sizeof(SECURITY_DESCRIPTOR_RELATIVE),
+ mode, isdir, usid, gsid);
+ if (((int)sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz + gsidsz) <= newattrsz) {
+ /* append usid and gsid */
+ memcpy(&newattr[sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz], usid, usidsz);
+ memcpy(&newattr[sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz], gsid, gsidsz);
+ /* positions of ACL, USID and GSID into header */
+ pnhead->owner =
+ cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz);
+ pnhead->group =
+ cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE)
+ + aclsz + usidsz);
+ pnhead->sacl = const_cpu_to_le32(0);
+ pnhead->dacl =
+ const_cpu_to_le32(sizeof(SECURITY_DESCRIPTOR_RELATIVE));
+ } else {
+ /* hope error was detected before overflowing */
+ free(newattr);
+ newattr = (char*)NULL;
+ ntfs_log_error("Security descriptor is longer than expected\n");
+ errno = EIO;
+ }
+ } else
+ errno = ENOMEM;
+ return (newattr);
+}
+
+/*
+ * Create a mode_t permission set
+ * from owner, group and world grants as represented in ACEs
+ */
+
+static int merge_permissions(BOOL isdir,
+ le32 owner, le32 group, le32 world, le32 special)
+
+{
+ int perm;
+
+ perm = 0;
+ /* build owner permission */
+ if (owner) {
+ if (isdir) {
+ /* exec if any of list, traverse */
+ if (owner & DIR_GEXEC)
+ perm |= S_IXUSR;
+ /* write if any of addfile, adddir, delchild */
+ if (owner & DIR_GWRITE)
+ perm |= S_IWUSR;
+ /* read if any of list */
+ if (owner & DIR_GREAD)
+ perm |= S_IRUSR;
+ } else {
+ /* exec if execute or generic execute */
+ if (owner & FILE_GEXEC)
+ perm |= S_IXUSR;
+ /* write if any of writedata or generic write */
+ if (owner & FILE_GWRITE)
+ perm |= S_IWUSR;
+ /* read if any of readdata or generic read */
+ if (owner & FILE_GREAD)
+ perm |= S_IRUSR;
+ }
+ }
+ /* build group permission */
+ if (group) {
+ if (isdir) {
+ /* exec if any of list, traverse */
+ if (group & DIR_GEXEC)
+ perm |= S_IXGRP;
+ /* write if any of addfile, adddir, delchild */
+ if (group & DIR_GWRITE)
+ perm |= S_IWGRP;
+ /* read if any of list */
+ if (group & DIR_GREAD)
+ perm |= S_IRGRP;
+ } else {
+ /* exec if execute */
+ if (group & FILE_GEXEC)
+ perm |= S_IXGRP;
+ /* write if any of writedata, appenddata */
+ if (group & FILE_GWRITE)
+ perm |= S_IWGRP;
+ /* read if any of readdata */
+ if (group & FILE_GREAD)
+ perm |= S_IRGRP;
+ }
+ }
+ /* build world permission */
+ if (world) {
+ if (isdir) {
+ /* exec if any of list, traverse */
+ if (world & DIR_GEXEC)
+ perm |= S_IXOTH;
+ /* write if any of addfile, adddir, delchild */
+ if (world & DIR_GWRITE)
+ perm |= S_IWOTH;
+ /* read if any of list */
+ if (world & DIR_GREAD)
+ perm |= S_IROTH;
+ } else {
+ /* exec if execute */
+ if (world & FILE_GEXEC)
+ perm |= S_IXOTH;
+ /* write if any of writedata, appenddata */
+ if (world & FILE_GWRITE)
+ perm |= S_IWOTH;
+ /* read if any of readdata */
+ if (world & FILE_GREAD)
+ perm |= S_IROTH;
+ }
+ }
+ /* build special permission flags */
+ if (special) {
+ if (special & FILE_APPEND_DATA)
+ perm |= S_ISUID;
+ if (special & FILE_WRITE_DATA)
+ perm |= S_ISGID;
+ if (special & FILE_READ_DATA)
+ perm |= S_ISVTX;
+ }
+ return (perm);
+}
+
+#if POSIXACLS
+
+/*
+ * Normalize a Posix ACL either from a sorted raw set of
+ * access ACEs or default ACEs
+ * (standard case : different owner, group and administrator)
+ */
+
+static int norm_std_permissions_posix(struct POSIX_SECURITY *posix_desc,
+ BOOL groupowns, int start, int count, int target)
+{
+ int j,k;
+ s32 id;
+ u16 tag;
+ u16 tagsset;
+ struct POSIX_ACE *pxace;
+ mode_t grantgrps;
+ mode_t grantwrld;
+ mode_t denywrld;
+ mode_t allow;
+ mode_t deny;
+ mode_t perms;
+ mode_t mode;
+
+ mode = 0;
+ tagsset = 0;
+ /*
+ * Determine what is granted to some group or world
+ * Also get denials to world which are meant to prevent
+ * execution flags to be inherited by plain files
+ */
+ pxace = posix_desc->acl.ace;
+ grantgrps = 0;
+ grantwrld = 0;
+ denywrld = 0;
+ for (j=start; j<(start + count); j++) {
+ if (pxace[j].perms & POSIX_PERM_DENIAL) {
+ /* deny world exec unless for default */
+ if ((pxace[j].tag == POSIX_ACL_OTHER)
+ && !start)
+ denywrld = pxace[j].perms;
+ } else {
+ switch (pxace[j].tag) {
+ case POSIX_ACL_GROUP_OBJ :
+ grantgrps |= pxace[j].perms;
+ break;
+ case POSIX_ACL_GROUP :
+ if (pxace[j].id)
+ grantgrps |= pxace[j].perms;
+ break;
+ case POSIX_ACL_OTHER :
+ grantwrld = pxace[j].perms;
+ break;
+ default :
+ break;
+ }
+ }
+ }
+ /*
+ * Collect groups of ACEs related to the same id
+ * and determine what is granted and what is denied.
+ * It is important the ACEs have been sorted
+ */
+ j = start;
+ k = target;
+ while (j < (start + count)) {
+ tag = pxace[j].tag;
+ id = pxace[j].id;
+ if (pxace[j].perms & POSIX_PERM_DENIAL) {
+ deny = pxace[j].perms | denywrld;
+ allow = 0;
+ } else {
+ deny = denywrld;
+ allow = pxace[j].perms;
+ }
+ j++;
+ while ((j < (start + count))
+ && (pxace[j].tag == tag)
+ && (pxace[j].id == id)) {
+ if (pxace[j].perms & POSIX_PERM_DENIAL)
+ deny |= pxace[j].perms;
+ else
+ allow |= pxace[j].perms;
+ j++;
+ }
+ /*
+ * Build the permissions equivalent to grants and denials
+ */
+ if (groupowns) {
+ if (tag == POSIX_ACL_MASK)
+ perms = ~deny;
+ else
+ perms = allow & ~deny;
+ } else
+ switch (tag) {
+ case POSIX_ACL_USER_OBJ :
+ perms = (allow | grantgrps | grantwrld) & ~deny;
+ break;
+ case POSIX_ACL_USER :
+ if (id)
+ perms = (allow | grantgrps | grantwrld)
+ & ~deny;
+ else
+ perms = allow;
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ perms = (allow | grantwrld) & ~deny;
+ break;
+ case POSIX_ACL_GROUP :
+ if (id)
+ perms = (allow | grantwrld) & ~deny;
+ else
+ perms = allow;
+ break;
+ case POSIX_ACL_MASK :
+ perms = ~deny;
+ break;
+ default :
+ perms = allow & ~deny;
+ break;
+ }
+ /*
+ * Store into a Posix ACE
+ */
+ if (tag != POSIX_ACL_SPECIAL) {
+ pxace[k].tag = tag;
+ pxace[k].id = id;
+ pxace[k].perms = perms
+ & (POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X);
+ tagsset |= tag;
+ k++;
+ }
+ switch (tag) {
+ case POSIX_ACL_USER_OBJ :
+ mode |= ((perms & 7) << 6);
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ case POSIX_ACL_MASK :
+ mode = (mode & 07707) | ((perms & 7) << 3);
+ break;
+ case POSIX_ACL_OTHER :
+ mode |= perms & 7;
+ break;
+ case POSIX_ACL_SPECIAL :
+ mode |= (perms & (S_ISVTX | S_ISUID | S_ISGID));
+ break;
+ default :
+ break;
+ }
+ }
+ if (!start) { /* not satisfactory */
+ posix_desc->mode = mode;
+ posix_desc->tagsset = tagsset;
+ }
+ return (k - target);
+}
+
+#endif /* POSIXACLS */
+
+/*
+ * Interpret an ACL and extract meaningful grants
+ * (standard case : different owner, group and administrator)
+ */
+
+static int build_std_permissions(const char *securattr,
+ const SID *usid, const SID *gsid, BOOL isdir)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ int offdacl;
+ int offace;
+ int acecnt;
+ int nace;
+ BOOL noown;
+ le32 special;
+ le32 allowown, allowgrp, allowall;
+ le32 denyown, denygrp, denyall;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ pacl = (const ACL*)&securattr[offdacl];
+ special = const_cpu_to_le32(0);
+ allowown = allowgrp = allowall = const_cpu_to_le32(0);
+ denyown = denygrp = denyall = const_cpu_to_le32(0);
+ noown = TRUE;
+ if (offdacl) {
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ } else {
+ acecnt = 0;
+ offace = 0;
+ }
+ for (nace = 0; nace < acecnt; nace++) {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if (!(pace->flags & INHERIT_ONLY_ACE)) {
+ if (ntfs_same_sid(usid, &pace->sid)
+ || ntfs_same_sid(ownersid, &pace->sid)) {
+ noown = FALSE;
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowown |= pace->mask;
+ else if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denyown |= pace->mask;
+ } else
+ if (ntfs_same_sid(gsid, &pace->sid)
+ && !(pace->mask & WRITE_OWNER)) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowgrp |= pace->mask;
+ else if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denygrp |= pace->mask;
+ } else
+ if (is_world_sid((const SID*)&pace->sid)) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowall |= pace->mask;
+ else
+ if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denyall |= pace->mask;
+ } else
+ if ((ntfs_same_sid((const SID*)&pace->sid,nullsid))
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE))
+ special |= pace->mask;
+ }
+ offace += le16_to_cpu(pace->size);
+ }
+ /*
+ * No indication about owner's rights : grant basic rights
+ * This happens for files created by Windows in directories
+ * created by Linux and owned by root, because Windows
+ * merges the admin ACEs
+ */
+ if (noown)
+ allowown = (FILE_READ_DATA | FILE_WRITE_DATA | FILE_EXECUTE);
+ /*
+ * Add to owner rights granted to group or world
+ * unless denied personaly, and add to group rights
+ * granted to world unless denied specifically
+ */
+ allowown |= (allowgrp | allowall);
+ allowgrp |= allowall;
+ return (merge_permissions(isdir,
+ allowown & ~(denyown | denyall),
+ allowgrp & ~(denygrp | denyall),
+ allowall & ~denyall,
+ special));
+}
+
+/*
+ * Interpret an ACL and extract meaningful grants
+ * (special case : owner and group are the same,
+ * and not administrator)
+ */
+
+static int build_owngrp_permissions(const char *securattr,
+ const SID *usid, BOOL isdir)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ int offdacl;
+ int offace;
+ int acecnt;
+ int nace;
+ le32 special;
+ BOOL grppresent;
+ le32 allowown, allowgrp, allowall;
+ le32 denyown, denygrp, denyall;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ pacl = (const ACL*)&securattr[offdacl];
+ special = const_cpu_to_le32(0);
+ allowown = allowgrp = allowall = const_cpu_to_le32(0);
+ denyown = denygrp = denyall = const_cpu_to_le32(0);
+ grppresent = FALSE;
+ if (offdacl) {
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ } else
+ acecnt = 0;
+ for (nace = 0; nace < acecnt; nace++) {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if (!(pace->flags & INHERIT_ONLY_ACE)) {
+ if ((ntfs_same_sid(usid, &pace->sid)
+ || ntfs_same_sid(ownersid, &pace->sid))
+ && (pace->mask & WRITE_OWNER)) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowown |= pace->mask;
+ } else
+ if (ntfs_same_sid(usid, &pace->sid)
+ && (!(pace->mask & WRITE_OWNER))) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE) {
+ allowgrp |= pace->mask;
+ grppresent = TRUE;
+ }
+ } else
+ if (is_world_sid((const SID*)&pace->sid)) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowall |= pace->mask;
+ else
+ if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denyall |= pace->mask;
+ } else
+ if ((ntfs_same_sid((const SID*)&pace->sid,nullsid))
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE))
+ special |= pace->mask;
+ }
+ offace += le16_to_cpu(pace->size);
+ }
+ if (!grppresent)
+ allowgrp = allowall;
+ return (merge_permissions(isdir,
+ allowown & ~(denyown | denyall),
+ allowgrp & ~(denygrp | denyall),
+ allowall & ~denyall,
+ special));
+}
+
+#if POSIXACLS
+
+/*
+ * Normalize a Posix ACL either from a sorted raw set of
+ * access ACEs or default ACEs
+ * (special case : owner or/and group is administrator)
+ */
+
+static int norm_ownadmin_permissions_posix(struct POSIX_SECURITY *posix_desc,
+ int start, int count, int target)
+{
+ int j,k;
+ s32 id;
+ u16 tag;
+ u16 tagsset;
+ struct POSIX_ACE *pxace;
+ int acccnt;
+ mode_t denywrld;
+ mode_t allow;
+ mode_t deny;
+ mode_t perms;
+ mode_t mode;
+
+ mode = 0;
+ pxace = posix_desc->acl.ace;
+ acccnt = posix_desc->acccnt;
+ tagsset = 0;
+ denywrld = 0;
+ /*
+ * Get denials to world which are meant to prevent
+ * execution flags to be inherited by plain files
+ */
+ for (j=start; j<(start + count); j++) {
+ if (pxace[j].perms & POSIX_PERM_DENIAL) {
+ /* deny world exec not for default */
+ if ((pxace[j].tag == POSIX_ACL_OTHER)
+ && !start)
+ denywrld = pxace[j].perms;
+ }
+ }
+ /*
+ * Collect groups of ACEs related to the same id
+ * and determine what is granted (denials are ignored)
+ * It is important the ACEs have been sorted
+ */
+ j = start;
+ k = target;
+ deny = 0;
+ while (j < (start + count)) {
+ allow = 0;
+ tag = pxace[j].tag;
+ id = pxace[j].id;
+ if (tag == POSIX_ACL_MASK) {
+ deny = pxace[j].perms;
+ j++;
+ while ((j < (start + count))
+ && (pxace[j].tag == POSIX_ACL_MASK))
+ j++;
+ } else {
+ if (!(pxace[j].perms & POSIX_PERM_DENIAL))
+ allow = pxace[j].perms;
+ j++;
+ while ((j < (start + count))
+ && (pxace[j].tag == tag)
+ && (pxace[j].id == id)) {
+ if (!(pxace[j].perms & POSIX_PERM_DENIAL))
+ allow |= pxace[j].perms;
+ j++;
+ }
+ }
+
+ /*
+ * Store the grants into a Posix ACE
+ */
+ if (tag == POSIX_ACL_MASK)
+ perms = ~deny;
+ else
+ perms = allow & ~denywrld;
+ if (tag != POSIX_ACL_SPECIAL) {
+ pxace[k].tag = tag;
+ pxace[k].id = id;
+ pxace[k].perms = perms
+ & (POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X);
+ tagsset |= tag;
+ k++;
+ }
+ switch (tag) {
+ case POSIX_ACL_USER_OBJ :
+ mode |= ((perms & 7) << 6);
+ break;
+ case POSIX_ACL_GROUP_OBJ :
+ case POSIX_ACL_MASK :
+ mode = (mode & 07707) | ((perms & 7) << 3);
+ break;
+ case POSIX_ACL_OTHER :
+ mode |= perms & 7;
+ break;
+ case POSIX_ACL_SPECIAL :
+ mode |= perms & (S_ISVTX | S_ISUID | S_ISGID);
+ break;
+ default :
+ break;
+ }
+ }
+ if (!start) { /* not satisfactory */
+ posix_desc->mode = mode;
+ posix_desc->tagsset = tagsset;
+ }
+ return (k - target);
+}
+
+#endif /* POSIXACLS */
+
+/*
+ * Interpret an ACL and extract meaningful grants
+ * (special case : owner or/and group is administrator)
+ */
+
+
+static int build_ownadmin_permissions(const char *securattr,
+ const SID *usid, const SID *gsid, BOOL isdir)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ int offdacl;
+ int offace;
+ int acecnt;
+ int nace;
+ BOOL firstapply;
+ int isforeign;
+ le32 special;
+ le32 allowown, allowgrp, allowall;
+ le32 denyown, denygrp, denyall;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ pacl = (const ACL*)&securattr[offdacl];
+ special = const_cpu_to_le32(0);
+ allowown = allowgrp = allowall = const_cpu_to_le32(0);
+ denyown = denygrp = denyall = const_cpu_to_le32(0);
+ if (offdacl) {
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ } else {
+ acecnt = 0;
+ offace = 0;
+ }
+ firstapply = TRUE;
+ isforeign = 3;
+ for (nace = 0; nace < acecnt; nace++) {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if (!(pace->flags & INHERIT_ONLY_ACE)
+ && !(~pace->mask & (ROOT_OWNER_UNMARK | ROOT_GROUP_UNMARK))) {
+ if ((ntfs_same_sid(usid, &pace->sid)
+ || ntfs_same_sid(ownersid, &pace->sid))
+ && (((pace->mask & WRITE_OWNER) && firstapply))) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE) {
+ allowown |= pace->mask;
+ isforeign &= ~1;
+ } else
+ if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denyown |= pace->mask;
+ } else
+ if (ntfs_same_sid(gsid, &pace->sid)
+ && (!(pace->mask & WRITE_OWNER))) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE) {
+ allowgrp |= pace->mask;
+ isforeign &= ~2;
+ } else
+ if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denygrp |= pace->mask;
+ } else if (is_world_sid((const SID*)&pace->sid)) {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ allowall |= pace->mask;
+ else
+ if (pace->type == ACCESS_DENIED_ACE_TYPE)
+ denyall |= pace->mask;
+ }
+ firstapply = FALSE;
+ } else
+ if (!(pace->flags & INHERIT_ONLY_ACE))
+ if ((ntfs_same_sid((const SID*)&pace->sid,nullsid))
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE))
+ special |= pace->mask;
+ offace += le16_to_cpu(pace->size);
+ }
+ if (isforeign) {
+ allowown |= (allowgrp | allowall);
+ allowgrp |= allowall;
+ }
+ return (merge_permissions(isdir,
+ allowown & ~(denyown | denyall),
+ allowgrp & ~(denygrp | denyall),
+ allowall & ~denyall,
+ special));
+}
+
+#if OWNERFROMACL
+
+/*
+ * Define the owner of a file as the first user allowed
+ * to change the owner, instead of the user defined as owner.
+ *
+ * This produces better approximations for files written by a
+ * Windows user in an inheritable directory owned by another user,
+ * as the access rights are inheritable but the ownership is not.
+ *
+ * An important case is the directories "Documents and Settings/user"
+ * which the users must have access to, though Windows considers them
+ * as owned by administrator.
+ */
+
+const SID *ntfs_acl_owner(const char *securattr)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const SID *usid;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ int offdacl;
+ int offace;
+ int acecnt;
+ int nace;
+ BOOL found;
+
+ found = FALSE;
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ if (offdacl) {
+ pacl = (const ACL*)&securattr[offdacl];
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ nace = 0;
+ do {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if ((pace->mask & WRITE_OWNER)
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ && ntfs_is_user_sid(&pace->sid))
+ found = TRUE;
+ offace += le16_to_cpu(pace->size);
+ } while (!found && (++nace < acecnt));
+ }
+ if (found)
+ usid = &pace->sid;
+ else
+ usid = (const SID*)&securattr[le32_to_cpu(phead->owner)];
+ return (usid);
+}
+
+#else
+
+/*
+ * Special case for files owned by administrator with full
+ * access granted to a mapped user : consider this user as the tenant
+ * of the file.
+ *
+ * This situation cannot be represented with Linux concepts and can
+ * only be found for files or directories created by Windows.
+ * Typical situation : directory "Documents and Settings/user" which
+ * is on the path to user's files and must be given access to user
+ * only.
+ *
+ * Check file is owned by administrator and no user has rights before
+ * calling.
+ * Returns the uid of tenant or zero if none
+ */
+
+
+static uid_t find_tenant(struct MAPPING *const mapping[],
+ const char *securattr)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ int offdacl;
+ int offace;
+ int acecnt;
+ int nace;
+ uid_t tid;
+ uid_t xid;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ pacl = (const ACL*)&securattr[offdacl];
+ tid = 0;
+ if (offdacl) {
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ } else
+ acecnt = 0;
+ for (nace = 0; nace < acecnt; nace++) {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if ((pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ && (pace->mask & DIR_WRITE)) {
+ xid = NTFS_FIND_USER(mapping[MAPUSERS], &pace->sid);
+ if (xid) tid = xid;
+ }
+ offace += le16_to_cpu(pace->size);
+ }
+ return (tid);
+}
+
+#endif /* OWNERFROMACL */
+
+#if POSIXACLS
+
+/*
+ * Build Posix permissions from an ACL
+ * returns a pointer to the requested permissions
+ * or a null pointer (with errno set) if there is a problem
+ *
+ * If the NTFS ACL was created according to our rules, the retrieved
+ * Posix ACL should be the exact ACL which was set. However if
+ * the NTFS ACL was built by a different tool, the result could
+ * be a a poor approximation of what was expected
+ */
+
+struct POSIX_SECURITY *ntfs_build_permissions_posix(
+ struct MAPPING *const mapping[],
+ const char *securattr,
+ const SID *usid, const SID *gsid, BOOL isdir)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ struct POSIX_SECURITY *pxdesc;
+ const ACL *pacl;
+ const ACCESS_ALLOWED_ACE *pace;
+ struct POSIX_ACE *pxace;
+ struct {
+ uid_t prevuid;
+ gid_t prevgid;
+ int groupmasks;
+ s16 tagsset;
+ BOOL gotowner;
+ BOOL gotownermask;
+ BOOL gotgroup;
+ mode_t permswrld;
+ } ctx[2], *pctx;
+ int offdacl;
+ int offace;
+ int alloccnt;
+ int acecnt;
+ uid_t uid;
+ gid_t gid;
+ int i,j;
+ int k,l;
+ BOOL ignore;
+ BOOL adminowns;
+ BOOL groupowns;
+ BOOL firstinh;
+ BOOL genericinh;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ offdacl = le32_to_cpu(phead->dacl);
+ if (offdacl) {
+ pacl = (const ACL*)&securattr[offdacl];
+ acecnt = le16_to_cpu(pacl->ace_count);
+ offace = offdacl + sizeof(ACL);
+ } else {
+ acecnt = 0;
+ offace = 0;
+ }
+ adminowns = FALSE;
+ groupowns = ntfs_same_sid(gsid,usid);
+ firstinh = FALSE;
+ genericinh = FALSE;
+ /*
+ * Build a raw posix security descriptor
+ * by just translating permissions and ids
+ * Add 2 to the count of ACE to be able to insert
+ * a group ACE later in access and default ACLs
+ * and add 2 more to be able to insert ACEs for owner
+ * and 2 more for other
+ */
+ alloccnt = acecnt + 6;
+ pxdesc = (struct POSIX_SECURITY*)malloc(
+ sizeof(struct POSIX_SECURITY)
+ + alloccnt*sizeof(struct POSIX_ACE));
+ k = 0;
+ l = alloccnt;
+ for (i=0; i<2; i++) {
+ pctx = &ctx[i];
+ pctx->permswrld = 0;
+ pctx->prevuid = -1;
+ pctx->prevgid = -1;
+ pctx->groupmasks = 0;
+ pctx->tagsset = 0;
+ pctx->gotowner = FALSE;
+ pctx->gotgroup = FALSE;
+ pctx->gotownermask = FALSE;
+ }
+ for (j=0; j<acecnt; j++) {
+ pace = (const ACCESS_ALLOWED_ACE*)&securattr[offace];
+ if (pace->flags & INHERIT_ONLY_ACE) {
+ pxace = &pxdesc->acl.ace[l - 1];
+ pctx = &ctx[1];
+ } else {
+ pxace = &pxdesc->acl.ace[k];
+ pctx = &ctx[0];
+ }
+ ignore = FALSE;
+ /*
+ * grants for root as a designated user or group
+ */
+ if ((~pace->mask & (ROOT_OWNER_UNMARK | ROOT_GROUP_UNMARK))
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ && ntfs_same_sid(&pace->sid, adminsid)) {
+ pxace->tag = (pace->mask & ROOT_OWNER_UNMARK ? POSIX_ACL_GROUP : POSIX_ACL_USER);
+ pxace->id = 0;
+ if ((pace->mask & (GENERIC_ALL | WRITE_OWNER))
+ && (pace->flags & INHERIT_ONLY_ACE))
+ ignore = genericinh = TRUE;
+ } else
+ if (ntfs_same_sid(usid, &pace->sid)) {
+ pxace->id = -1;
+ /*
+ * Owner has no write-owner right :
+ * a group was defined same as owner
+ * or admin was owner or group :
+ * denials are meant to owner
+ * and grants are meant to group
+ */
+ if (!(pace->mask & (WRITE_OWNER | GENERIC_ALL))
+ && (pace->type == ACCESS_ALLOWED_ACE_TYPE)) {
+ if (ntfs_same_sid(gsid,usid)) {
+ pxace->tag = POSIX_ACL_GROUP_OBJ;
+ pxace->id = -1;
+ } else {
+ if (ntfs_same_sid(&pace->sid,usid))
+ groupowns = TRUE;
+ gid = NTFS_FIND_GROUP(mapping[MAPGROUPS],&pace->sid);
+ if (gid) {
+ pxace->tag = POSIX_ACL_GROUP;
+ pxace->id = gid;
+ pctx->prevgid = gid;
+ } else {
+ uid = NTFS_FIND_USER(mapping[MAPUSERS],&pace->sid);
+ if (uid) {
+ pxace->tag = POSIX_ACL_USER;
+ pxace->id = uid;
+ } else
+ ignore = TRUE;
+ }
+ }
+ } else {
+ /*
+ * when group owns, late denials for owner
+ * mean group mask
+ */
+ if ((pace->type == ACCESS_DENIED_ACE_TYPE)
+ && (pace->mask & WRITE_OWNER)) {
+ pxace->tag = POSIX_ACL_MASK;
+ pctx->gotownermask = TRUE;
+ if (pctx->gotowner)
+ pctx->groupmasks++;
+ } else {
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ pctx->gotowner = TRUE;
+ if (pctx->gotownermask && !pctx->gotowner) {
+ uid = NTFS_FIND_USER(mapping[MAPUSERS],&pace->sid);
+ pxace->id = uid;
+ pxace->tag = POSIX_ACL_USER;
+ } else
+ pxace->tag = POSIX_ACL_USER_OBJ;
+ /* system ignored, and admin */
+ /* ignored at first position */
+ if (pace->flags & INHERIT_ONLY_ACE) {
+ if ((firstinh && ntfs_same_sid(&pace->sid,adminsid))
+ || ntfs_same_sid(&pace->sid,systemsid))
+ ignore = TRUE;
+ if (!firstinh) {
+ firstinh = TRUE;
+ }
+ } else {
+ if ((adminowns && ntfs_same_sid(&pace->sid,adminsid))
+ || ntfs_same_sid(&pace->sid,systemsid))
+ ignore = TRUE;
+ if (ntfs_same_sid(usid,adminsid))
+ adminowns = TRUE;
+ }
+ }
+ }
+ } else if (ntfs_same_sid(gsid, &pace->sid)) {
+ if ((pace->type == ACCESS_DENIED_ACE_TYPE)
+ && (pace->mask & WRITE_OWNER)) {
+ pxace->tag = POSIX_ACL_MASK;
+ pxace->id = -1;
+ if (pctx->gotowner)
+ pctx->groupmasks++;
+ } else {
+ if (pctx->gotgroup || (pctx->groupmasks > 1)) {
+ gid = NTFS_FIND_GROUP(mapping[MAPGROUPS],&pace->sid);
+ if (gid) {
+ pxace->id = gid;
+ pxace->tag = POSIX_ACL_GROUP;
+ pctx->prevgid = gid;
+ } else
+ ignore = TRUE;
+ } else {
+ pxace->id = -1;
+ pxace->tag = POSIX_ACL_GROUP_OBJ;
+ if (pace->type == ACCESS_ALLOWED_ACE_TYPE)
+ pctx->gotgroup = TRUE;
+ }
+
+ if (ntfs_same_sid(gsid,adminsid)
+ || ntfs_same_sid(gsid,systemsid)) {
+ if (pace->mask & (WRITE_OWNER | GENERIC_ALL))
+ ignore = TRUE;
+ if (ntfs_same_sid(gsid,adminsid))
+ adminowns = TRUE;
+ else
+ genericinh = ignore;
+ }
+ }
+ } else if (is_world_sid((const SID*)&pace->sid)) {
+ pxace->id = -1;
+ pxace->tag = POSIX_ACL_OTHER;
+ if ((pace->type == ACCESS_DENIED_ACE_TYPE)
+ && (pace->flags & INHERIT_ONLY_ACE))
+ ignore = TRUE;
+ } else if (ntfs_same_sid((const SID*)&pace->sid,nullsid)) {
+ pxace->id = -1;
+ pxace->tag = POSIX_ACL_SPECIAL;
+ } else {
+ uid = NTFS_FIND_USER(mapping[MAPUSERS],&pace->sid);
+ if (uid) {
+ if ((pace->type == ACCESS_DENIED_ACE_TYPE)
+ && (pace->mask & WRITE_OWNER)
+ && (pctx->prevuid != uid)) {
+ pxace->id = -1;
+ pxace->tag = POSIX_ACL_MASK;
+ } else {
+ pxace->id = uid;
+ pxace->tag = POSIX_ACL_USER;
+ }
+ pctx->prevuid = uid;
+ } else {
+ gid = NTFS_FIND_GROUP(mapping[MAPGROUPS],&pace->sid);
+ if (gid) {
+ if ((pace->type == ACCESS_DENIED_ACE_TYPE)
+ && (pace->mask & WRITE_OWNER)
+ && (pctx->prevgid != gid)) {
+ pxace->tag = POSIX_ACL_MASK;
+ pctx->groupmasks++;
+ } else {
+ pxace->tag = POSIX_ACL_GROUP;
+ }
+ pxace->id = gid;
+ pctx->prevgid = gid;
+ } else {
+ /*
+ * do not grant rights to unknown
+ * people and do not define root as a
+ * designated user or group
+ */
+ ignore = TRUE;
+ }
+ }
+ }
+ if (!ignore) {
+ pxace->perms = 0;
+ /* specific decoding for vtx/uid/gid */
+ if (pxace->tag == POSIX_ACL_SPECIAL) {
+ if (pace->mask & FILE_APPEND_DATA)
+ pxace->perms |= S_ISUID;
+ if (pace->mask & FILE_WRITE_DATA)
+ pxace->perms |= S_ISGID;
+ if (pace->mask & FILE_READ_DATA)
+ pxace->perms |= S_ISVTX;
+ } else
+ if (isdir) {
+ if (pace->mask & DIR_GEXEC)
+ pxace->perms |= POSIX_PERM_X;
+ if (pace->mask & DIR_GWRITE)
+ pxace->perms |= POSIX_PERM_W;
+ if (pace->mask & DIR_GREAD)
+ pxace->perms |= POSIX_PERM_R;
+ if ((pace->mask & GENERIC_ALL)
+ && (pace->flags & INHERIT_ONLY_ACE))
+ pxace->perms |= POSIX_PERM_X
+ | POSIX_PERM_W
+ | POSIX_PERM_R;
+ } else {
+ if (pace->mask & FILE_GEXEC)
+ pxace->perms |= POSIX_PERM_X;
+ if (pace->mask & FILE_GWRITE)
+ pxace->perms |= POSIX_PERM_W;
+ if (pace->mask & FILE_GREAD)
+ pxace->perms |= POSIX_PERM_R;
+ }
+
+ if (pace->type != ACCESS_ALLOWED_ACE_TYPE)
+ pxace->perms |= POSIX_PERM_DENIAL;
+ else
+ if (pxace->tag == POSIX_ACL_OTHER)
+ pctx->permswrld = pxace->perms;
+ pctx->tagsset |= pxace->tag;
+ if (pace->flags & INHERIT_ONLY_ACE) {
+ l--;
+ } else {
+ k++;
+ }
+ }
+ offace += le16_to_cpu(pace->size);
+ }
+ /*
+ * Create world perms if none (both lists)
+ */
+ for (i=0; i<2; i++)
+ if ((genericinh || !i)
+ && !(ctx[i].tagsset & POSIX_ACL_OTHER)) {
+ if (i)
+ pxace = &pxdesc->acl.ace[--l];
+ else
+ pxace = &pxdesc->acl.ace[k++];
+ pxace->tag = POSIX_ACL_OTHER;
+ pxace->id = -1;
+ pxace->perms = 0;
+ ctx[i].tagsset |= POSIX_ACL_OTHER;
+ ctx[i].permswrld = 0;
+ }
+ /*
+ * Set basic owner perms if none (both lists)
+ * This happens for files created by Windows in directories
+ * created by Linux and owned by root, because Windows
+ * merges the admin ACEs
+ */
+ for (i=0; i<2; i++)
+ if (!(ctx[i].tagsset & POSIX_ACL_USER_OBJ)
+ && (ctx[i].tagsset & POSIX_ACL_OTHER)) {
+ if (i)
+ pxace = &pxdesc->acl.ace[--l];
+ else
+ pxace = &pxdesc->acl.ace[k++];
+ pxace->tag = POSIX_ACL_USER_OBJ;
+ pxace->id = -1;
+ pxace->perms = POSIX_PERM_R | POSIX_PERM_W | POSIX_PERM_X;
+ ctx[i].tagsset |= POSIX_ACL_USER_OBJ;
+ }
+ /*
+ * Duplicate world perms as group_obj perms if none
+ */
+ for (i=0; i<2; i++)
+ if ((ctx[i].tagsset & POSIX_ACL_OTHER)
+ && !(ctx[i].tagsset & POSIX_ACL_GROUP_OBJ)) {
+ if (i)
+ pxace = &pxdesc->acl.ace[--l];
+ else
+ pxace = &pxdesc->acl.ace[k++];
+ pxace->tag = POSIX_ACL_GROUP_OBJ;
+ pxace->id = -1;
+ pxace->perms = ctx[i].permswrld;
+ ctx[i].tagsset |= POSIX_ACL_GROUP_OBJ;
+ }
+ /*
+ * Also duplicate world perms as group perms if they
+ * were converted to mask and not followed by a group entry
+ */
+ if (ctx[0].groupmasks) {
+ for (j=k-2; j>=0; j--) {
+ if ((pxdesc->acl.ace[j].tag == POSIX_ACL_MASK)
+ && (pxdesc->acl.ace[j].id != -1)
+ && ((pxdesc->acl.ace[j+1].tag != POSIX_ACL_GROUP)
+ || (pxdesc->acl.ace[j+1].id
+ != pxdesc->acl.ace[j].id))) {
+ pxace = &pxdesc->acl.ace[k];
+ pxace->tag = POSIX_ACL_GROUP;
+ pxace->id = pxdesc->acl.ace[j].id;
+ pxace->perms = ctx[0].permswrld;
+ ctx[0].tagsset |= POSIX_ACL_GROUP;
+ k++;
+ }
+ if (pxdesc->acl.ace[j].tag == POSIX_ACL_MASK)
+ pxdesc->acl.ace[j].id = -1;
+ }
+ }
+ if (ctx[1].groupmasks) {
+ for (j=l; j<(alloccnt-1); j++) {
+ if ((pxdesc->acl.ace[j].tag == POSIX_ACL_MASK)
+ && (pxdesc->acl.ace[j].id != -1)
+ && ((pxdesc->acl.ace[j+1].tag != POSIX_ACL_GROUP)
+ || (pxdesc->acl.ace[j+1].id
+ != pxdesc->acl.ace[j].id))) {
+ pxace = &pxdesc->acl.ace[l - 1];
+ pxace->tag = POSIX_ACL_GROUP;
+ pxace->id = pxdesc->acl.ace[j].id;
+ pxace->perms = ctx[1].permswrld;
+ ctx[1].tagsset |= POSIX_ACL_GROUP;
+ l--;
+ }
+ if (pxdesc->acl.ace[j].tag == POSIX_ACL_MASK)
+ pxdesc->acl.ace[j].id = -1;
+ }
+ }
+
+ /*
+ * Insert default mask if none present and
+ * there are designated users or groups
+ * (the space for it has not beed used)
+ */
+ for (i=0; i<2; i++)
+ if ((ctx[i].tagsset & (POSIX_ACL_USER | POSIX_ACL_GROUP))
+ && !(ctx[i].tagsset & POSIX_ACL_MASK)) {
+ if (i)
+ pxace = &pxdesc->acl.ace[--l];
+ else
+ pxace = &pxdesc->acl.ace[k++];
+ pxace->tag = POSIX_ACL_MASK;
+ pxace->id = -1;
+ pxace->perms = POSIX_PERM_DENIAL;
+ ctx[i].tagsset |= POSIX_ACL_MASK;
+ }
+
+ if (k > l) {
+ ntfs_log_error("Posix descriptor is longer than expected\n");
+ errno = EIO;
+ free(pxdesc);
+ pxdesc = (struct POSIX_SECURITY*)NULL;
+ } else {
+ pxdesc->acccnt = k;
+ pxdesc->defcnt = alloccnt - l;
+ pxdesc->firstdef = l;
+ pxdesc->tagsset = ctx[0].tagsset;
+ pxdesc->acl.version = POSIX_VERSION;
+ pxdesc->acl.flags = 0;
+ pxdesc->acl.filler = 0;
+ ntfs_sort_posix(pxdesc);
+ if (adminowns) {
+ k = norm_ownadmin_permissions_posix(pxdesc,
+ 0, pxdesc->acccnt, 0);
+ pxdesc->acccnt = k;
+ l = norm_ownadmin_permissions_posix(pxdesc,
+ pxdesc->firstdef, pxdesc->defcnt, k);
+ pxdesc->firstdef = k;
+ pxdesc->defcnt = l;
+ } else {
+ k = norm_std_permissions_posix(pxdesc,groupowns,
+ 0, pxdesc->acccnt, 0);
+ pxdesc->acccnt = k;
+ l = norm_std_permissions_posix(pxdesc,groupowns,
+ pxdesc->firstdef, pxdesc->defcnt, k);
+ pxdesc->firstdef = k;
+ pxdesc->defcnt = l;
+ }
+ }
+ if (pxdesc && !ntfs_valid_posix(pxdesc)) {
+ ntfs_log_error("Invalid Posix descriptor built\n");
+ errno = EIO;
+ free(pxdesc);
+ pxdesc = (struct POSIX_SECURITY*)NULL;
+ }
+ return (pxdesc);
+}
+
+#endif /* POSIXACLS */
+
+/*
+ * Build unix-style (mode_t) permissions from an ACL
+ * returns the requested permissions
+ * or a negative result (with errno set) if there is a problem
+ */
+
+int ntfs_build_permissions(const char *securattr,
+ const SID *usid, const SID *gsid, BOOL isdir)
+{
+ const SECURITY_DESCRIPTOR_RELATIVE *phead;
+ int perm;
+ BOOL adminowns;
+ BOOL groupowns;
+
+ phead = (const SECURITY_DESCRIPTOR_RELATIVE*)securattr;
+ adminowns = ntfs_same_sid(usid,adminsid)
+ || ntfs_same_sid(gsid,adminsid);
+ groupowns = !adminowns && ntfs_same_sid(gsid,usid);
+ if (adminowns)
+ perm = build_ownadmin_permissions(securattr, usid, gsid, isdir);
+ else
+ if (groupowns)
+ perm = build_owngrp_permissions(securattr, usid, isdir);
+ else
+ perm = build_std_permissions(securattr, usid, gsid, isdir);
+ return (perm);
+}
+
+/*
+ * The following must be in some library...
+ */
+
+static unsigned long atoul(const char *p)
+{ /* must be somewhere ! */
+ unsigned long v;
+
+ v = 0;
+ while ((*p >= '0') && (*p <= '9'))
+ v = v * 10 + (*p++) - '0';
+ return (v);
+}
+
+/*
+ * Build an internal representation of a SID
+ * Returns a copy in allocated memory if it succeeds
+ * The SID is checked to be a valid user one.
+ */
+
+static SID *encodesid(const char *sidstr)
+{
+ SID *sid;
+ int cnt;
+ BIGSID bigsid;
+ SID *bsid;
+ u32 auth;
+ const char *p;
+
+ sid = (SID*) NULL;
+ if (!strncmp(sidstr, "S-1-", 4)) {
+ bsid = (SID*)&bigsid;
+ bsid->revision = SID_REVISION;
+ p = &sidstr[4];
+ auth = atoul(p);
+ bsid->identifier_authority.high_part = const_cpu_to_be16(0);
+ bsid->identifier_authority.low_part = cpu_to_be32(auth);
+ cnt = 0;
+ p = strchr(p, '-');
+ while (p && (cnt < 8)) {
+ p++;
+ auth = atoul(p);
+ bsid->sub_authority[cnt] = cpu_to_le32(auth);
+ p = strchr(p, '-');
+ cnt++;
+ }
+ bsid->sub_authority_count = cnt;
+ if ((cnt > 0) && ntfs_valid_sid(bsid) && ntfs_is_user_sid(bsid)) {
+ sid = (SID*) ntfs_malloc(4 * cnt + 8);
+ if (sid)
+ memcpy(sid, bsid, 4 * cnt + 8);
+ }
+ }
+ return (sid);
+}
+
+/*
+ * Early logging before the logs are redirected
+ *
+ * (not quite satisfactory : this appears before the ntfs-g banner,
+ * and with a different pid)
+ */
+
+static void log_early_error(const char *format, ...)
+ __attribute__((format(printf, 1, 2)));
+
+static void log_early_error(const char *format, ...)
+{
+ va_list args;
+
+ va_start(args, format);
+#ifdef HAVE_SYSLOG_H
+ openlog("ntfs-3g", LOG_PID, LOG_USER);
+ ntfs_log_handler_syslog(NULL, NULL, 0,
+ NTFS_LOG_LEVEL_ERROR, NULL,
+ format, args);
+#else
+ vfprintf(stderr,format,args);
+#endif
+ va_end(args);
+}
+
+
+/*
+ * Get a single mapping item from buffer
+ *
+ * Always reads a full line, truncating long lines
+ * Refills buffer when exhausted
+ * Returns pointer to item, or NULL when there is no more
+ */
+
+static struct MAPLIST *getmappingitem(FILEREADER reader, void *fileid,
+ off_t *poffs, char *buf, int *psrc, s64 *psize)
+{
+ int src;
+ int dst;
+ char *p;
+ char *q;
+ char *pu;
+ char *pg;
+ int gotend;
+ struct MAPLIST *item;
+
+ src = *psrc;
+ dst = 0;
+ /* allocate and get a full line */
+ item = (struct MAPLIST*)ntfs_malloc(sizeof(struct MAPLIST));
+ if (item) {
+ do {
+ gotend = 0;
+ while ((src < *psize)
+ && (buf[src] != '\n')) {
+ if (dst < LINESZ)
+ item->maptext[dst++] = buf[src];
+ src++;
+ }
+ if (src >= *psize) {
+ *poffs += *psize;
+ *psize = reader(fileid, buf, (size_t)BUFSZ, *poffs);
+ src = 0;
+ } else {
+ gotend = 1;
+ src++;
+ item->maptext[dst] = '\0';
+ dst = 0;
+ }
+ } while (*psize && ((item->maptext[0] == '#') || !gotend));
+ if (gotend) {
+ pu = pg = (char*)NULL;
+ /* decompose into uid, gid and sid */
+ p = item->maptext;
+ item->uidstr = item->maptext;
+ item->gidstr = strchr(item->uidstr, ':');
+ if (item->gidstr) {
+ pu = item->gidstr++;
+ item->sidstr = strchr(item->gidstr, ':');
+ if (item->sidstr) {
+ pg = item->sidstr++;
+ q = strchr(item->sidstr, ':');
+ if (q) *q = 0;
+ }
+ }
+ if (pu && pg)
+ *pu = *pg = '\0';
+ else {
+ log_early_error("Bad mapping item \"%s\"\n",
+ item->maptext);
+ free(item);
+ item = (struct MAPLIST*)NULL;
+ }
+ } else {
+ free(item); /* free unused item */
+ item = (struct MAPLIST*)NULL;
+ }
+ }
+ *psrc = src;
+ return (item);
+}
+
+/*
+ * Read user mapping file and split into their attribute.
+ * Parameters are kept as text in a chained list until logins
+ * are converted to uid.
+ * Returns the head of list, if any
+ *
+ * If an absolute path is provided, the mapping file is assumed
+ * to be located in another mounted file system, and plain read()
+ * are used to get its contents.
+ * If a relative path is provided, the mapping file is assumed
+ * to be located on the current file system, and internal IO
+ * have to be used since we are still mounting and we have not
+ * entered the fuse loop yet.
+ */
+
+struct MAPLIST *ntfs_read_mapping(FILEREADER reader, void *fileid)
+{
+ char buf[BUFSZ];
+ struct MAPLIST *item;
+ struct MAPLIST *firstitem;
+ struct MAPLIST *lastitem;
+ int src;
+ off_t offs;
+ s64 size;
+
+ firstitem = (struct MAPLIST*)NULL;
+ lastitem = (struct MAPLIST*)NULL;
+ offs = 0;
+ size = reader(fileid, buf, (size_t)BUFSZ, (off_t)0);
+ if (size > 0) {
+ src = 0;
+ do {
+ item = getmappingitem(reader, fileid, &offs,
+ buf, &src, &size);
+ if (item) {
+ item->next = (struct MAPLIST*)NULL;
+ if (lastitem)
+ lastitem->next = item;
+ else
+ firstitem = item;
+ lastitem = item;
+ }
+ } while (item);
+ }
+ return (firstitem);
+}
+
+/*
+ * Free memory used to store the user mapping
+ * The only purpose is to facilitate the detection of memory leaks
+ */
+
+void ntfs_free_mapping(struct MAPPING *mapping[])
+{
+ struct MAPPING *user;
+ struct MAPPING *group;
+
+ /* free user mappings */
+ while (mapping[MAPUSERS]) {
+ user = mapping[MAPUSERS];
+ /* do not free SIDs used for group mappings */
+ group = mapping[MAPGROUPS];
+ while (group && (group->sid != user->sid))
+ group = group->next;
+ if (!group)
+ free(user->sid);
+ /* free group list if any */
+ if (user->grcnt)
+ free(user->groups);
+ /* unchain item and free */
+ mapping[MAPUSERS] = user->next;
+ free(user);
+ }
+ /* free group mappings */
+ while (mapping[MAPGROUPS]) {
+ group = mapping[MAPGROUPS];
+ free(group->sid);
+ /* unchain item and free */
+ mapping[MAPGROUPS] = group->next;
+ free(group);
+ }
+}
+
+
+/*
+ * Build the user mapping list
+ * user identification may be given in symbolic or numeric format
+ *
+ * ! Note ! : does getpwnam() read /etc/passwd or some other file ?
+ * if so there is a possible recursion into fuse if this
+ * file is on NTFS, and fuse is not recursion safe.
+ */
+
+struct MAPPING *ntfs_do_user_mapping(struct MAPLIST *firstitem)
+{
+ struct MAPLIST *item;
+ struct MAPPING *firstmapping;
+ struct MAPPING *lastmapping;
+ struct MAPPING *mapping;
+ struct passwd *pwd;
+ SID *sid;
+ int uid;
+
+ firstmapping = (struct MAPPING*)NULL;
+ lastmapping = (struct MAPPING*)NULL;
+ for (item = firstitem; item; item = item->next) {
+ if ((item->uidstr[0] >= '0') && (item->uidstr[0] <= '9'))
+ uid = atoi(item->uidstr);
+ else {
+ uid = 0;
+ if (item->uidstr[0]) {
+ pwd = getpwnam(item->uidstr);
+ if (pwd)
+ uid = pwd->pw_uid;
+ else
+ log_early_error("Invalid user \"%s\"\n",
+ item->uidstr);
+ }
+ }
+ /*
+ * Records with no uid and no gid are inserted
+ * to define the implicit mapping pattern
+ */
+ if (uid
+ || (!item->uidstr[0] && !item->gidstr[0])) {
+ sid = encodesid(item->sidstr);
+ if (sid && !item->uidstr[0] && !item->gidstr[0]
+ && !ntfs_valid_pattern(sid)) {
+ ntfs_log_error("Bad implicit SID pattern %s\n",
+ item->sidstr);
+ sid = (SID*)NULL;
+ }
+ if (sid) {
+ mapping =
+ (struct MAPPING*)
+ ntfs_malloc(sizeof(struct MAPPING));
+ if (mapping) {
+ mapping->sid = sid;
+ mapping->xid = uid;
+ mapping->grcnt = 0;
+ mapping->next = (struct MAPPING*)NULL;
+ if (lastmapping)
+ lastmapping->next = mapping;
+ else
+ firstmapping = mapping;
+ lastmapping = mapping;
+ }
+ }
+ }
+ }
+ return (firstmapping);
+}
+
+/*
+ * Build the group mapping list
+ * group identification may be given in symbolic or numeric format
+ *
+ * gid not associated to a uid are processed first in order
+ * to favour real groups
+ *
+ * ! Note ! : does getgrnam() read /etc/group or some other file ?
+ * if so there is a possible recursion into fuse if this
+ * file is on NTFS, and fuse is not recursion safe.
+ */
+
+struct MAPPING *ntfs_do_group_mapping(struct MAPLIST *firstitem)
+{
+ struct MAPLIST *item;
+ struct MAPPING *firstmapping;
+ struct MAPPING *lastmapping;
+ struct MAPPING *mapping;
+ struct group *grp;
+ BOOL secondstep;
+ BOOL ok;
+ int step;
+ SID *sid;
+ int gid;
+
+ firstmapping = (struct MAPPING*)NULL;
+ lastmapping = (struct MAPPING*)NULL;
+ for (step=1; step<=2; step++) {
+ for (item = firstitem; item; item = item->next) {
+ secondstep = (item->uidstr[0] != '\0')
+ || !item->gidstr[0];
+ ok = (step == 1 ? !secondstep : secondstep);
+ if ((item->gidstr[0] >= '0')
+ && (item->gidstr[0] <= '9'))
+ gid = atoi(item->gidstr);
+ else {
+ gid = 0;
+ if (item->gidstr[0]) {
+ grp = getgrnam(item->gidstr);
+ if (grp)
+ gid = grp->gr_gid;
+ else
+ log_early_error("Invalid group \"%s\"\n",
+ item->gidstr);
+ }
+ }
+ /*
+ * Records with no uid and no gid are inserted in the
+ * second step to define the implicit mapping pattern
+ */
+ if (ok
+ && (gid
+ || (!item->uidstr[0] && !item->gidstr[0]))) {
+ sid = encodesid(item->sidstr);
+ if (sid && !item->uidstr[0] && !item->gidstr[0]
+ && !ntfs_valid_pattern(sid)) {
+ /* error already logged */
+ sid = (SID*)NULL;
+ }
+ if (sid) {
+ mapping = (struct MAPPING*)
+ ntfs_malloc(sizeof(struct MAPPING));
+ if (mapping) {
+ mapping->sid = sid;
+ mapping->xid = gid;
+ mapping->grcnt = 0;
+ mapping->next = (struct MAPPING*)NULL;
+ if (lastmapping)
+ lastmapping->next = mapping;
+ else
+ firstmapping = mapping;
+ lastmapping = mapping;
+ }
+ }
+ }
+ }
+ }
+ return (firstmapping);
+}